Characterization of an intact two-component high-affinity nitrate transporter from Arabidopsis roots


For correspondence (fax +1 604 822 6089; e-mail


AtNRT2.1, a polypeptide of the Arabidopsis thaliana two-component inducible high-affinity nitrate transport system (IHATS), is located within the plasma membrane. The monomeric form of AtNRT2.1 has been reported to be the most abundant form, and was suggested to be the form that is active in nitrate transport. Here we have used immunological and transient protoplast expression methods to demonstrate that an intact two-component complex of AtNRT2.1 and AtNAR2.1 (AtNRT3.1) is localized in the plasma membrane. A. thaliana mutants lacking AtNAR2.1 have virtually no IHATS capacity and exhibit extremely poor growth on low nitrate as the sole source of nitrogen. Near-normal growth and nitrate transport in the mutant were restored by transformation with myc-tagged AtNAR2.1 cDNA. Membrane fractions from roots of the restored myc-tagged line were solubilized in 1.5% dodecyl-β-maltoside and partially purified in the first dimension by blue native gel electrophoresis. Using anti-NRT2.1 antibodies, an oligomeric polypeptide (approximate molecular mass 150 kDa) was identified, but monomeric AtNRT2.1 was absent. This oligomer was also observed in the wild-type, and was resolved, using SDS–PAGE for the second dimension, into two polypeptides with molecular masses of approximately 48 and 26 kDa, corresponding to AtNRT2.1 and myc-tagged AtNAR2.1, respectively. This result, together with the finding that the oligomer is absent from NRT2.1 or NAR2.1 mutants, suggests that this complex, rather than monomeric AtNRT2.1, is the form that is active in IHATS nitrate transport. The molecular mass of the intact oligomer suggests that the functional unit for high-affinity nitrate influx may be a tetramer consisting of two subunits each of AtNRT2.1 and AtNAR2.1.


Inducible high-affinity nitrate transport (IHATS) in Arabidopsis thaliana has been shown to require expression of two genes, namely AtNRT2.1 and AtNAR2.1 (AtNRT3.1) (Filleur et al., 2001; Orsel et al., 2004, 2006; Okamoto et al., 2006; Li et al., 2007). IHATS in T-DNA mutants disrupted in AtNRT2.1 was reduced by approximately 70% (Filleur et al., 2001; Li et al., 2007), while disruption of AtNAR2.1 caused IHATS to be reduced by as much as 95% (Okamoto et al., 2006; Orsel et al., 2006). Furthermore, growth of mutants disrupted in AtNRT2.1, AtNRT2.1/AtNRT2.2 or AtNAR2.1 is severely restricted at low external nitrate concentration (Okamoto et al., 2006; Orsel et al., 2006; Li et al., 2007). By contrast, low-affinity nitrate transport (LATS), encoded by AtNRT1.1 (Tsay et al., 1993), appears not to require simultaneous expression of AtNAR2.1, as AtNAR2.1 mutants exhibited normal LATS function (Okamoto et al., 2006). The requirement for simultaneous expression of two gene products in order to sustain high-affinity nitrate transport was first demonstrated genetically in Chlamydomonas reinhardtii, in which the capacity for high-affinity transport by CrNRT2.1 or CrNRT2.2 was lost in mutants lacking CrNAR2 (Quesada et al., 1994). Further support for a two-component high-affinity nitrate influx was provided by the finding that nitrate transport in Xenopus oocytes requires co-expression of CrNRT2 and CrNAR2 (Zhou et al., 2000). Likewise, only when both of the barley homologues (HvNRT2.1 and HvNAR2.3) were co-expressed in Xenopus oocytes did nitrate transport occur (Tong et al., 2005). The positive results reported for a yeast two-hybrid split-ubiquitin assay further suggest that the functional high-affinity nitrate transporter may involve an intimate interaction between AtNRT2.1 and AtNAR2.1 (reviewed by Glass, 2009). Nevertheless, despite this indirect evidence, no higher-order complex consisting of AtNRT2.1 and AtNAR2.1 has so far been demonstrated. Although Wirth et al. (2007) identified a high-molecular-weight polypeptide at approximately 120 kDa in wild-type (WT) plants using NRT2.1 antibodies, the continued strong expression of this polypeptide in the Atnar2.1 knockout mutant eliminated it as a possible candidate for the putative higher-order complex of AtNRT2.1 and AtNAR2.1.

Using a combination of a GFP fusion and immunological methods, it was demonstrated that AtNRT2.1 is mainly localized in the plasma membranes (PMs) of root cortical and epidermal cells (Chopin et al., 2007; Wirth et al., 2007). Wirth et al. (2007) suggested that several forms of this polypeptide, i.e. a 45 kDa monomer and higher-molecular-weight forms, co-exist. They found that the monomeric form was the most abundant, and suggested that it was the form involved in nitrate transport. The authors attempted to cross-link AtNRT2.1 with its partner AtNAR2.1 by treating microsomal fractions with 1% formaldehyde prior to extraction and SDS–PAGE, but did not detect a putative complex consisting of the two participants in high-affinity nitrate influx, and concluded that NAR2.1 was not a part of the high-molecular-mass polypeptide of approximately 120 kDa. In a recent study of the two-component system in barley (Ishikawa et al., 2009), it was demonstrated that both HvNRT2.1 and HvNAR2.3 are localized in the PM, and the authors suggested that the C-terminus of HvNRT2.1 may be involved in binding to the central region of HvNAR2.3. Indeed, an earlier paper (Kawachi et al., 2006) reported that a point mutation in the middle region of AtNAR2.1 resulted in loss of HATS activity.

We used the Atnar2.1-2 T-DNA mutant described previously (Okamoto et al., 2006) as the recipient of a myc-tagged AtNAR2.1 cDNA. The transformed mutant recovered near-normal growth and high-affinity 13NO3 influx. This line, as well as WT and other lines, was used for isolation of microsomal and plasma membrane fractions. Using blue native PAGE (BN–PAGE) and immunological methods, we successfully identified a PM oligomer (molecular mass approximately 150 kDa) that was resolved into the component monomers AtNRT2.1 and AtNAR2.1 (molecular masses approximately 48 and 26 kDa, respectively) by SDS–PAGE in the second dimension. Localization of the two-component complex consisting of AtNRT2.1 and AtNAR2.1 was confirmed by in vivo transient expression of split YFP-labelled AtNRT2.1 and AtNAR2.1 in Arabidopsis protoplasts.


Restoration of phenotype in the Atnar2.1-35S:NAR2.1-myc mutant

Figure 1(a,b) shows the growth of the wild-type (WT), Atnar2.1 mutant and the two myc lines on agar medium containing 0.25 or 10 mm KNO3. The mutant showed virtually no growth at low nitrate, but grew similarly to WT when grown on high nitrate or NH4NO3 (Okamoto et al., 2006). Transformation with a myc-tagged NAR2.1 cDNA (AtNAR2.1-myc) restored growth of the mutant to near-WT rates when grown on low-nitrate medium (Figure 1 and Figure S1). The dry weight of roots also returned to WT values, and shoot growth increased from 6% of the WT value in the mutant to approximately 70% of the WT value in the restored lines (Table 1). As previously reported (Okamoto et al., 2006), 13NO3 influx was reduced to 6% of the WT value in the Atnar2.1 mutant when plants were grown hydroponically as described below. Following transformation with AtNAR2.1-myc, the rates of 13NO3 influx increased to approximately 60 and 70% of the WT flux in the two lines (Table 2).

Figure 1.

 Growth of various Arabidopsis lines (WT Ws, Atnar2.1 mutant and Atnar2.1-35S:NRT2.1-myc lines) on half-strength MS agar containing (a) 0.25 mm KNO3 and (b) 10 mm KNO3.

Table 1.   Dry weight of plants hydroponically grown for 5 weeks at 250 μm KNO3
GenotypeRoot weight (mg)Percentage of WT valueShoot weight (mg)Percentage of WT value
  1. Values are means ± SE of ten replicates. *Difference significant at < 0.05 compared with WT.

Wild-type (Ws)1.40 ± 0.211009.97 ± 1.10100
Atnar2.10.53 ± 0.07*380.64 ± 0.05*6
Atnar2.1-NAR2.1-myc21.35 ± 0.10966.12 ± 0.33*61
Atnar2.1-NAR2.1-myc41.22 ± 0.14876.76 ± 0.64*68
Table 2. 13NO3 influx into roots of WT, Atnar2.1 mutant and Atnar2.1-35S:NAR2.1-myc lines after induction with 1 mm KNO3 for 6 h
GenotypeInflux (μmol g FW−1 h−1)Percentage of WT flux
  1. Influx was measured using 0.1 mm KNO3 to determine IHATS activity. Values are means ± SE of six replicates.

Wild-type (Ws)5.58 ± 0.37100
Atnar2.10.36 ± 0.066
Atnar2.1-NAR2.1-myc23.26 ± 0.1358
Atnar2.1-NAR2.1-myc43.72 ± 0.1967

Absence of AtNRT2.1 in various mutants

Lines of various mutant plants disrupted in AtNRT2.1 only (Li et al., 2007), in both AtNRT2.1 and AtNRT2.2 (Filleur et al., 2001; Li et al., 2007) or in AtNAR2.1 only (Okamoto et al., 2006) were grown, and microsomal fractions were prepared from roots of these plants after 4 weeks of hydroponic growth on medium containing 1 mm NH4NO3 followed by 1 week without nitrogen and 6 h induction with 1 mm KNO3. This treatment maximizes expression of AtNRT2.1, AtNAR2.1 and inducible high-affinity 13NO3 influx (Okamoto et al., 2006). The microsomal fractions were subjected to SDS–PAGE, and Western blots were prepared and probed using anti-AtNRT2.1 antibody. Figure 2 shows that AtNRT2.1 was present in WT microsomal fractions but was absent from fractions isolated from the Atnrt2.1 mutant, the double mutant (Atnrt2.1-nrt2.2) and the Atnar2.1 mutant.

Figure 2.

 Western blot of microsomal fractions from roots of various Arabidopsis lines after SDS–PAGE, probed using anti-NRT2.1 antibody.
Lane 1, WT (Col); lane 2, Atnrt2.1 mutant; lane 3, Atnrt2.1-nrt2.2 double mutant; lane 4, Atnar2.1 mutant.

Plasma membrane localization of AtNRT2.1 and AtNAR2.1

Roots of the rescued A. thaliana Atnar2.1 mutant (Atnar2.1-35S:NAR2.1-myc line) containing myc-tagged AtNAR2.1 and WT AtNRT2.1 were used as the source of plant material for isolation of membrane fractions. Plants were grown hydroponically in the same manner as for flux determination (described above). The roots of the rescued Atnar2.1-35S:NAR2.1-myc4 line were separated from shoot tissue and used to isolate microsomes. This microsomal preparation was subjected to sucrose-gradient centrifugation followed by SDS–PAGE. Western blots were then probed using anti-myc antibodies. Figure 3 shows that the anti-myc antibodies recognized the presence of myc-tagged AtNAR2.1 in microsomal-, plasma membrane- and endoplasmic reticulum (ER)/Golgi-enriched fractions, but not in a tonoplast-enriched fraction. We also used two-phase partitioning in PEG/dextran to obtain a highly purified plasma membrane (PM) fraction. The identity and purity of this preparation was evaluated by determining ATPase activity in the presence and absence of 1.0 mm vanadate, a specific inhibitor of PM H+-ATPase activity, and by use of specific antibodies. 1 mm sodium vanadate inhibited ATPase activity of microsomes and PM by 70 and 80%, respectively (Table 3). The microsomal and PEG/dextran-generated PM fractions were subjected to SDS–PAGE followed by Western blotting, and then probed using polyclonal antibodies against ER- and tonoplast-specific proteins to evaluate the extent of contamination by these membranes, and against AtNRT2.1 and AtNAR2.1 to verify their presence in the purified PM fraction (Figure S2). The almost complete absence of a reaction to anti-V-PPase (tonoplast marker: vacuolar H+ pyrophosphatase) and anti-Bip (ER marker: luminal-binding protein) indicated a lack of significant contamination by these membranes, and the reactions to anti-AtNRT2.1 and anti-myc confirm the presence of these polypeptides in the purified PM fraction. Figure S2 also confirms molecular masses of 48 and 26 kDa for these polypeptides.

Figure 3.

 Western blot of various membrane-enriched fractions from roots of the Atnar2.1-35S:NAR2.1-myc4 line separated by sucrose gradient centrifugation, followed by SDS–PAGE and probing with anti-myc antibody to localize myc-tagged AtNAR2.1.
Lane 1, microsomes; lane 2, tonoplast; lane 3, plasma membrane; lane 4, endoplasmic reticulum/Golgi.

Table 3.   ATPase activity of microsomes and PEG/dextran-purified PM
Membrane fractionATPase activity (μmol Pi mg−1 h−1)Percentage inhibition
Without vanadateWith 1 mm vanadate
  1. Values are means ± SE of six replicates.

Microsomes19.43 ± 1.695.93 ± 0.7170
PM45.56 ± 0.838.95 ± 0.3280

Identification of the intact AtNRT2.1/AtNAR2.1 complex

The Atnar2.1-35S:NAR2.1-myc4 line, Atnar2.1, Atnrt2.1 and Atnrt2.1-nrt2.2 mutants and WT plants were grown under standard conditions (see above), and roots of these plants were used for isolation of microsomal membranes and PEG/dextran-generated PM fractions. These were solubilized in 1.5% dodecyl maltoside and subjected to BN–PAGE in the first dimension. Resulting Western blots were probed with anti-NRT2.1 antibody. Figure 4(a,b) shows that the anti-NRT2.1 antibody gave a positive reaction with a protein complex of molecular mass approximately 150 kDa that was present in both microsomes and purified PM fractions. The molecular mass of this complex was estimated as described previously (Yamaoka et al., 1993; Yamaoka, 1998). To ensure that the complex of approximately 150 kDa was not an artefact resulting from the presence of the 35S promoter in 35S:NAR2.1-myc, the above procedures were repeated using WT-derived membranes. Figure 4(c) shows that this complex was also present in membranes isolated from WT roots. No free NRT2.1 was detected in any of the BN–PAGE preparations shown in Figure 4. However, this complex was absent from samples prepared using roots of the Atnar2.1 knockout mutant described by Okamoto et al. (2006) and the Atnrt2.1 and Atnrt2.1-nrt2.2 mutant lines (Li et al., 2007), as shown in Figure 4(d). The 150 kDa complexes from the WT and the rescued line gave no reaction to anti-myc antibody.

Figure 4.

 Western blots after BN–PAGE and probing with anti-NRT2.1 antibody.
(a) Microsomal fraction from roots of the Atnar2.1-35S:NAR2.1-myc4 line.
(b) PEG/dextran-purified plasma membrane fraction from roots of the Atnar2.1-35S:NRT2.1-myc4 line.
(c) Microsomal fraction from roots of WT plants.
(d) Comparison of microsomal fraction from roots of WT Col (lane 1), Atnar2.1 (lane 2), Atnrt2.1 (lane 3) and Atnrt2.1-nrt2.2 (lane 4).

The entire lane from the BN–PAGE of microsomes from the Atnar2.1-35S:NAR2.1-myc4 mutant was cut out and transferred to an SDS–polyacrylamide gel for second-dimension electrophoresis. After SDS–PAGE, the gel was used for Western blotting and probed first with anti-myc antibody. This antibody reacted with a polypeptide with a molecular mass of approximately 26 kDa (Figure 5a) derived from a region of the BN–PAGE gel corresponding to a molecular mass of approximately 150 kDa. After stripping, the PVDF membrane was probed with anti-AtNRT2.1 antibody, and AtNRT2.1 was detected at a molecular mass of approximately 48 kDa. Figure 5(b) shows that AtNRT2.1 was also derived from the region of the BN–PAGE gel corresponding to a molecular mass of approximately 150 kDa. Overlapping the two immunoblots (Figure 5c) clearly shows that both AtNRT2.1 and AtNAR2.1 are derived from the same region of the BN–PAGE gel. The above procedures were repeated using PEG/dextran-purified PM preparations with identical results to those shown in Figure 5. Thus, unlike the 120 kDa polypeptide reported by Wirth et al. (2007), our complex of AtNRT2.1 and AtNAR2.1 was separated by SDS treatment.

Figure 5.

 Separation of the native 150 kDa complex using SDS–PAGE in the second dimension.
(a) Western blot after SDS–PAGE of the blue native gel lane and probing with anti-myc antibody.
(b) After stripping the membrane, the same Western blot was probed using anti-NRT2.1 antibody.
(c) Overlap of the images in (a) and (b) demonstrates that the two polypeptides are derived from the same approximately 150 kDa region of the blue native gel.

The PM localization of the complex was also investigated by transient in vivo expression of split YFP-labelled AtNRT2.1 and AtNAR2.1 in WT Arabidopsis leaf protoplasts. Figure 6 shows bright-field and fluorescence images of protoplasts transformed with both AtNRT2.1–cEYFP and AtNAR2.1–nEYFP (Figure 6a) and protoplasts transformed with either AtNRT2.1–cEYFP and nEYFP or with AtNAR2.1–nEYFP and cEYFP as controls (Figure 6b,c). Only when both AtNRT2.1 and AtNAR2.1 were present was fluorescence localized to the PM.

Figure 6.

 Interaction of AtNRT2.1 and AtNAR2.1 in vivo.
Transient expression of split YFP constructs in Arabidopsis leaf protoplasts. Bright-field (left) and fluorescence (right) images. Protoplasts were transfected with (a) AtNRT2.1–cEYFP and AtNAR2.1–nEYFP, (b) AtNAR2.1–nEYFP and cEYFP, or (c) AtNRT2.1–cEYFP and nEYFP.


Inducible high-affinity nitrate influx depends upon coincident expression of two genes, namely NRT2.1 and NAR2.1, in several species, including C. reinhardtii (Quesada et al., 1994), A. thaliana (Filleur et al., 2001; Okamoto et al., 2006; Orsel et al., 2006; Li et al., 2007), H. vulgare (Tong et al., 2005), and rice, wheat and the moss Physcomitrella patens (reviewed by Glass, 2009). Disruption of AtNAR2.1 by a T-DNA insertion reduced IHATS 13NO3 influx to approximately 5% of WT values (Okamoto et al., 2006). Nevertheless, the transcript abundance of AtNRT2.1 was still relatively high, and it was subsequently demonstrated that the AtNRT2.1 polypeptide was absent in PM preparations from this mutant (Wirth et al., 2007). The authors concluded that AtNAR2.1 is essential for proper targeting of AtNRT2.1 to the PM, and also suggested that AtNAR2.1 might be necessary to stabilize AtNRT2.1.

We propose that the IHATS nitrate transporter is stable and functional only in the presently demonstrated oligomeric form. Our anti-NRT2.1 antibody recognized the presence of AtNRT2.1 in microsomal membranes and purified PM-enriched preparations from roots of WT plants and the Atnar2.1-35S:NAR2.1-myc line (Figures 2 and S2). The absence of a reaction to membranes from the Atnrt2.1 mutant (Figure 2) demonstrates that the antibody is specific for AtNRT2.1 and does not recognize AtNRT2.2. Likewise, the antibody did not recognize any of the other five NRT polypeptides in membranes from the roots of Atnrt2.1 or Atnrt2.1-nrt2.2 mutants. The absence of any reaction to membranes from the Atnar2.1 mutant (Figure 2) confirms that AtNRT2.1 is absent from the PM in this mutant (Wirth et al., 2007). Figure 2 confirms earlier observations that disruption of AtNAR2.1 in T-DNA mutants renders the mutant essentially incapable of growth on low concentrations of nitrate (Okamoto et al., 2006; Orsel et al., 2006), and our 13NO3 assay demonstrated that HATS activity was reduced to approximately 6% of WT values. The restoration of near-WT plant growth and HATS activity confirms that the Atnar2.1-35S:NAR2.1-myc4 line expresses functional AtNAR2.1 (Figures 1 and S1, Tables 1 and 2), although neither growth nor measured 13NO3 influx were completely restored, remaining at approximately 70% of WT values. The reasons for this incomplete recovery are unclear. Perhaps positional effects are responsible, as complete (100%) recovery was not observed for the transformed lines. Alternatively, the effect may be a consequence of several T-DNA insertions. Nevertheless, the presence of the 150 kDa complex in WT roots as well as in the ‘rescued’ line provides proof that this complex is the natural form of AtNRT2.1/AtNAR2.

The data shown in Figures 3 and 5 and Figure S2 confirm that AtNAR2.1 is localized in the PM, using PM-enriched fractions derived from sucrose-gradient centrifugation and preparations that were highly purified using PEG/dextran. The identity of the latter preparations was verified by the extent of vanadate-inhibitable ATPase activity (Table 3), and their purity was established by the absence of ER or tonoplast membranes (Figure S2). Thus both members of the two-component high-affinity nitrate transporter are localized in the PM of A. thaliana, as already demonstrated for barley (Ishikawa et al., 2009). Wirth et al. (2007) had previously shown that AtNRT2.1 is localized in the PM; we confirm that AtNAR2.1 is also localized in this membrane.

Microsome and PEG/dextran-purified PM preparations from the Atnar2.1-35S:NAR2.1-myc line subjected to BN–PAGE and probed with anti-AtNRT2.1 antibody showed a positive reaction with a complex of molecular mass approximately 150 kDa that was resolved into two polypeptides of molecular masses equal to approximately 48 and 26 kDa when the entire BN–PAGE lane was subjected to second-dimension SDS–PAGE. These polypeptides correspond to AtNRT2.1 and AtNAR2.1, respectively, based upon their reaction with the corresponding antibodies (Figures 4 and 5). The co-migration of these two polypeptides from the approximately 150 kDa region of the BN–PAGE gel establishes that these PM polypeptides originated from a single complex of approximately 150 kDa. This same 150 kDa complex was also present in the roots of WT plants (Figure 4c). The separation of these two polypeptides by SDS–PAGE demonstrates that the two polypeptides are bound by non-covalent linkages, possibly between the C-terminus of AtNRT2.1 and the central portion of AtNAR2.1, as suggested in a study of the barley system (Ishikawa et al., 2009). Interestingly, the absence of any free AtNRT2.1 in the BN–PAGE immunoblots (Figure 4) and the localization of NAR2.1 in the PM strongly suggest that AtNRT2.1 is present only as a complex with AtNAR2.1. These observations conflict with the suggestion (Wirth et al., 2007) that the role of NAR2.1 is in processing of NRT2.1 rather than permanent association with NRT2.1 in the PM.

In contrast to the study by Wirth et al. (2007), we did not detect anti-AtNRT2.1 reactive polypeptides at approximately 75 and 120 kDa. Nor were polypeptides of these molecular masses observed in the barley study (Ishikawa et al., 2009). More importantly, Wirth et al. (2007) concluded that NAR2.1 was not a part of their 120 kDa polypeptide because the polypeptide was still abundant in the Atnar2.1 mutant. If this polypeptide had contained AtNAR2.1 and AtNRT2.1, the 120 kDa polypeptide should have been separated by denaturing SDS treatment, as they were in our study (Figure 5). The absence of our 150 kDa complex from membranes isolated from roots of the Atnrt2.1, Atnrt2.1-nrt2.2 and Atnar2.1 mutants (Figure 4d) is consistent with the involvement of both AtNRT2.1 and AtNAR2.1 in the complex. The combined molecular mass of a single subunit each of AtNRT2.1 and AtNAR2.1 is approximately 74 kDa. Therefore, given the estimated molecular mass of approximately 150 kDa for the native complex, we suggest that the functional inducible high-affinity nitrate transporter may be a tetramer consisting of two subunits each of AtNRT2.1 and AtNAR2.1. The failure of the anti-myc antibody to recognize the AtNAR2.1 polypeptide in the complex of approximately 150 kDa that so clearly contains both AtNRT2.1 and AtNAR2.1 polypeptides suggests that the proposed two subunits of AtNRT2.1 may enclose the AtNAR2.1 subunits, making the myc peptide inaccessible to the antibody.

The results indicating reconstituted fluorescence (Figure 6) only when both AtNRT2.1 and AtNAR2.1 are transiently expressed in the leaf protoplasts confirm the intimate in vivo association between these polypeptides, and the subcellular pattern of fluorescence confirms the findings for the immunological methods (discussed above) with respect to the PM localization of the complex.

Although HATS in Arabidopsis depends upon expression of both AtNRT2.1 and AtNAR2.1, no equivalent to the AtNAR2.1 gene has been found in the fungus Aspergillus nidulans (S.E. Unkles, School of Biology, University of St Andrews, personal communication), and nitrate fluxes were generated when AnNRTA (the Aspergillus NRT2 homologue) alone was expressed in the Xenopus oocyte system (Zhou et al., 2000). By contrast, nitrate influx was obtained only when both CrNRT2.1 and CrNAR2 were co-injected into Xenopus oocytes. This was also the case for the barley homologues HvNRT2.1 and HvNRT2.3 (Tong et al., 2005). This difference between the fungal and plant systems may be related to the length of the central cytoplasmic loop between transmembrane regions 6 and 7 and/or to the long C-terminus in plants. In plants (including Chlamydomonas), the cytoplasmic loop is considerably shorter than that of Aspergillus and Hansenulla, e.g. the A. thaliana NRT2.1 loop consists of 21 amino acids compared to 95 amino acids in A. nidulans NRTA. It was suggested (Ishikawa et al., 2009) that the interaction between HvNRT2.1 and HvNAR2 in barley occurs via the C-terminus of HvNRT2.1 and the central region of HvNAR2.

In conclusion, the results of the present study are at variance with several of the conclusions of Wirth et al. (2007). In particular, we have observed:

  • (i) a unique PM complex made up of both AtNRT2.1 and AtNAR2.1 that is completely absent in knockout mutants Atnrt2.1 and Atnar2.1
  • (ii) complete absence of free AtNRT2.1 in PM preparations from membrane fractions solubilized with dodecyl β-maltoside, suggesting that AtNRT2.1 is only present in association with AtNAR2.1 and that the 150 kDa complex rather than the monomeric form of AtNRT2.1 is involved in high-affinity nitrate transport
  • (iii) an absence of higher-molecular-weight forms of AtNRT2.1 (at 75 or 120 kDa) when membranes were SDS-solubilized, as also indicated in the recent barley study by Ishikawa et al. (2009)
  • (iv) the presence of monomeric AtNRT2.1 only when SDS was used to solubilize membrane proteins (a condition that resulted in complete separation of AtNRT2.1 and AtNAR2.1)
  • (v) a molecular mass of approximately 150 kDa, suggesting that the observed AtNRT2.1/AtNAR2.1 complex is a tetramer consisting of two subunits each of AtNRT2.1 and AtNAR2.1.

Experimental procedures

Preparation of the Atnar2.1-35S:NAR2.1-myc lines

A myc-tagged NAR2.1 gene was cloned from Arabidopsis cDNA using a high-fidelity enzyme (Phusion®; Finnzymes, and the primers 5′-ATGGATCCATGGCGATCCAGAAGATCCTCTT-3′ (forward) and 5′-ATGAATTCTCAATTCAGATCCTCTTCTGAGATGAGTTTTTGTTCTTTGCTTTGCTCTATCTTGGCC-3′ (reverse). A modified binary pGreenII179 vector, containing a hygromycin resistance gene (Hellens et al., 2000a,b), was used to create the 35S:NAR2.1-myc construct. The Arabidopsis knockout Atnar2.1-2 line was transformed using the simplified Agrobacterium-mediated floral-dip method (Clough and Bent, 1998). T0 seed was subjected to selection on half-strength MS agar plates with 20 mg l−1 Hygromycin B (Invitrogen, Seed collected from T1 and T2 plants was used for all experiments.

Plant material and growth conditions

Arabidopsis plants of WT ecotypes Columbia-0 and Wassilewskija, and knockout mutant lines Atnrt2.1 (SALK_141712), Atnrt2.1-nrt2.2 (SALK_035429) and Atnar2.1-2 (Okamoto et al., 2006) were grown hydroponically under non-sterile conditions as described previously (Zhuo et al.,1999; Alonso et al., 2003; Okamoto et al., 2003). Three or four seeds were sown into 1.5 cm plastic cylinders filled with acid-washed sand and attached to floating StyrofoamTM (DOW, platforms. The platforms floated in plastic containers filled with 7 L of nutrient solution (1 mm KH2PO4, 0.5 mm MgSO4, 0.25 mm CaSO4, 20 μm Fe-EDTA, 25 μm H3BO3, 2 μm ZnSO4, 2 μm MnSO4, 0.5 μm CuSO4, 0.5 μm Na2MoO4 and 1 mm NH4NO3). Solutions were aerated continuously using aquarium stones, and the pH of solutions was maintained at approximately 6 by adding powdered CaCO3. Nutrient solutions were completely replaced weekly. Plants were grown for 4 weeks, and then deprived of nitrogen for the 5th week. To induce HATS (for protein expression and 13NO3 influx analysis), plants were then transferred to a solution containing 1 mm KNO3 for 6 h. Growth conditions in the growth room were 8 h light (100 μmol m−2 sec−1 at plant level) and 16 h dark, at corresponding temperatures of 24 and 20°C, respectively, and relative humidity of approximately 70%. For dry weight measurement, plants were grown in the same hydroponic nutrient solution described above, except that NH4NO3 was replaced by 0.25 mm KNO3. Roots and shoots were separated and dried at room temperature for 2 days.

For growth on MS agar plates, Arabidopsis seeds were sterilized in 1% bleach (plus 0.01% Tween-20) for 15 min, and left for 3 days in sterile water at 4°C for imbibition. Seeds were then sown on half-strength nitrogen-free MS salts, supplemented with 0.25 mm KNO3 or 10 mm KNO3. The plates were kept in a vertical position and plants were grown for 2 weeks under the same conditions as described above. Root length was measured using ImageJ and the NeuronJ plug-in (Meijering et al., 2004).

13NO3 influx measurements

Nitrate influx using 13NO3 was measured as described previously (Zhuo et al., 1999; Okamoto et al., 2003). The basic components of the solution for pre-treatment, influx and desorption were the same as those of the growth medium, except that 0.1 mm KNO3 replaced NH4NO3. Prior to measuring 13NO3 influx, plants were pre-treated for 5 min with solution containing 0.1 mm KNO3, and then transferred for 5 min into the influx solution, which was labelled with 13NO3. After the influx period, roots were desorbed using non-labelled solution (identical to pre-treatment solution) for 2 min to desorb the radioactive isotope from the apoplast. Gamma emission was measured using a MINAXI Auto-Gamma 5000 series gamma counter (Packard Instruments,

Microsome preparation

Arabidopsis roots were homogenized in homogenizing buffer consisting of 0.33 m sucrose, 5 mm EGTA, 2 mm salicylhydroxamic acid, 1 mm DTT, 1.5% soluble PVP, proteinase inhibitor and 25 mm Tris/Mes at pH 7.6. The homogenate was centrifuged at 10 000 g for 20 min, and the supernatant was then centrifuged at 100 000 g for 40 min. The pelleted microsomes were dispersed in resuspension buffer consisting of 0.33 m sucrose, 1 mm EDTA, 10 mm KCl, 1 mm DTT, protease inhibitor cocktail (Complete EDTA-free tablets, Roche, and 5 mm Tris/MES at pH 7.3, and centrifuged again at 100 000 g for 40 min. The pellet was again dispersed in resuspension buffer.

Isolation of plasma membranes

Plasma membranes were separated by two-phase partitioning as described by Santoni (2007). Microsomes were loaded into a two-phase system consisting of 6.3% w/w PEG 3350 (Sigma-Aldrich, and Dextran T500 (Pharmacia, in a final concentration of 0.33 mm sucrose, 3 mm KCl and 5 mm potassium phosphate (pH 7.8). The two phases were thoroughly mixed and centrifuged at 1500 g for 10 min. The top phase was removed and subjected to repartition by mixing with a new lower phase. After the second partitioning, the top phase was diluted in resuspension buffer and pelleted by centrifugation at 100 000 g for 40 min. The resulting plasma membrane pellets were resuspended in resuspension buffer and frozen at −80°C until required for further analysis.

Sucrose step-gradient fractionation

Solutions of various sucrose concentrations (15, 30, 34 and 45%) were prepared by solubilizing sucrose in sucrose gradient buffer consisting of 5 mm Tris/Mes (pH 7.3), 1 mm EDTA, 1 mm DTT, 10 mm KCl and protease inhibitor cocktail (Complete EDTA-free tablets, Roche). The 45% sucrose solution was carefully overlaied with the 38, 30 and 15% sucrose solutions. The microsome sample was layered over the 15% layer. The gradients were centrifuged at 80 000 g for 2 h. Bands formed at each interphase were carefully collected and diluted with sucrose gradient buffer and spun at 100 000 g for 40 min. The pellets were resuspended in resuspension buffer, and frozen at −80°C until further use.

Assay for plasma membrane H+-ATPase

Vanadate-sensitive, K+-stimulated Mg-ATPase activity (Leonard and Hodges, 1973) was determined by measuring the release of inorganic P (Ames, 1966). The reaction mixtures contained 3 mm ATP (Tris form), 3 mm MgSO4, 5 mm sodium azide, 1 mm sodium molybdate, 50 mm potassium nitrate, 50 mm potassium chloride, 0.2% Triton X-100, 2 mm EDTA and 250 mm sucrose in 30 mm Tris/Mes buffer (pH 6.5), with or without 1 mm sodium orthovanadate. Membrane protein (10–30 μg) was added to 0.45 ml reaction mixture to start the reaction, and the mixture was incubated at 36°C for 30 min. After incubation, the tubes were transferred to ice, 0.5 ml of ice-cold TCA/perchloric acid mixture (10% w/v TCA and 4% v/v perchloric acid) were added to stop the reaction, and samples were incubated on ice for a further 30 min. The precipitate formed was then pelleted by centrifugation at 10 000 g for 5 min, and 0.5 ml samples of the supernatant were transferred to clean test tubes. To each sample, 1 ml of Ames reagent (six parts 0.42% w/v ammonium molybdate in 1 N H2SO4 to one part 10% w/v ascorbic acid) was added, and the samples were incubated for 1 h at room temperature. The absorbance was then measured at 800 nm. A standard curve was prepared using KH2PO4.

Immunoblot analysis

For SDS immunoblotting analysis, proteins were separated by denaturing 12% SDS–PAGE followed by electrotransfer at 4°C onto a PVDF membrane (Hybond-P, Amersham, NRT2.1 was detected using anti-NRT2.1 antiserum produced by Alpha Diagnostic International ( against the following synthetic peptides (C)DLPDGNRATLEKAGE, (C)KNMHQGSLRFAENAK and (C)GRRVRSAATPPENTPNNV. The polyclonal antiserum was affinity-purified by Alpha Diagnostic International. The myc tag was detected using C-terminal myc antibody (Santa Cruz Biotechnology ( ER and tonoplast markers were detected using polyclonal anti-Bip and anti-V-PPase antibodies, respectively (both COSMO BIO, Immunodetection was performed using an ECL system kit (GE Healthcare,

BN–PAGE was performed as described previously (Schagger et al., 1994; Guo et al., 2005). An equal volume of resuspension buffer containing 3% w/v dodecyl β-d-maltoside was added to microsomes or PM suspension. After incubation at 4°C for 5 min, samples were combined with a one-tenth volume of 5% Serva Blue G in 100 mm BisTris/HCl (pH 7.0), 0.5 m 6-amino-N-caproic acid, 30% w/v glycerol, and applied to 1.5 mm thick 5–16% acrylamide gradient gels in a Hoefer SE 250 Mighty Small II (Hoefer, vertical electrophoresis unit operated at 4°C. For direct immunoblotting analysis, the lanes from BN–PAGE were cut out and equilibrated for 1 h in 1 × gel buffer (50 mm BisTris/HCl, 0.5 m 6-amino-N-caproic acid, pH 7.0) with 1% w/v SDS and 2.5% v/v β-mercaptoethanol. Then the samples were electrotransferred onto PVDF membranes for immunoblotting analysis as described above.

For separation in the second dimension using SDS–PAGE, the lanes from the first-dimension BN–PAGE were cut out and equilibrated for 1 h in SDS loading buffer and placed into a 12% acrylamide gel of the same thickness. Immunoblotting was performed using anti-myc antibody to detect the presence of myc-tagged NAR2.1. After NAR2.1 detection, the PVDF membrane was washed with stripping buffer (62.5 mm Tris/HCl pH 6.8, 2% SDS and 100 mmβ-mercaptoethanol) at 50°C for 30 min by shaking slowly. Membranes were then washed three times with a mixture of 10 mm Tris-HCl, 150 mm NaCl, and 0.05% Tween 20 at pH 7.5 for 10 min each. The washed PVDF membrane was used for immunodetection with anti-NRT2.1 antibody to detect NRT2.1. After immunoblotting analysis, the two films were overlaid so that the high-molecular-weight markers for native electrophoresis (GE Healthcare) lined up on both films.

Transient expression in protoplasts

NRT2.1 and NAR2.1 were tagged with halves of YFP using pSAT vectors (Citovsky et al., 2006). cDNA for AtNRT2.1 was fused in-frame to the C-terminal half of YFP in pSAT4A-cEYFP-N1 (XhoI/BamHI restriction sites). cDNA for AtNAR2.1 was fused in-frame to the N-terminal half of YFP in pSAT1A-nEYFP-N1 (XhoI/BamHI restriction sites). WT Arabidopsis leaf protoplasts were prepared and transformed with the constructs as described by Tiwari et al. (2006). Briefly, 1 g of leaves were cut into <1 mm strips and incubated in 1% w/v Cellulase (Onozuka R-10) and 0.25% w/v Macerozyme R-10 (Bio-world, for 90 min. Protoplasts were recovered by centrifugation at 180 g at room temperature for 3 min, resuspended at 3 × 105 protoplasts/ml, and transfected with 10 μg plasmid DNA by PEG-mediated transformation. After 18 h incubation in darkness, protoplasts were visualized using a spinning disk Perkin-Elmer ( UltraView VoX microscope (equipped with a Leica DMI6000, inverted microscope and a Hamamatsu 9100-02 CCD camera, and Volocity software (


The authors gratefully acknowledge financial support for this project in the form of Discovery Grant 0570 from the Natural Sciences and Engineering Research Council of Canada to A.D.M.G. and a University of British Columbia graduate fellowship to Z.K. The authors gratefully acknowledge the University of British Columbia (TRIUMF) cyclotron facility for provision of 13NO3. The authors would like to thank the Arabidopsis Biological Resource Center and the John Innes Centre for provision of pSAT and pGreen vectors, respectively, and the University of British Columbia BioImaging facility for assistance with fluorescence microscopy.