Identification and functional assay of the interaction motifs in the partner protein OsNAR2.1 of the two-component system for high-affinity nitrate transport

A partner protein, NAR2, is essential for high-affinity nitrate transport of the NRT2 protein in plants. However, the NAR2 motifs that interact with NRT2s for their plasma membrane (PM) localization and nitrate transporter activity have not been functionally characterized. In this study, OsNAR2.1 mutations with different carbon (C)-terminal deletions and nine different point mutations in the conserved regions of NAR2 homologs in plants were generated to explore the essential motifs involved in the interaction with OsNRT2.3a. Screening using the membrane yeast two-hybrid system and Xenopus oocytes for nitrogen-15 (15N) uptake demonstrated that either R100G or D109N point mutations impaired the OsNAR2.1 interaction with OsNRT2.3a. Western blotting and visualization using green fluorescent protein fused to either the N- or C-terminus of OsNAR2.1 indicated that OsNAR2.1 is expressed in both the PM and cytoplasm. The split-yellow fluorescent protein (YFP)/BiFC analyses indicated that OsNRT2.3a was targeted to the PM in the presence of OsNAR2.1, while either R100G or D109N mutation resulted in the loss of OsNRT2.3a-YFP signal in the PM. Based on these results, arginine 100 and aspartic acid 109 of the OsNAR2.1 protein are key amino acids in the interaction with OsNRT2.3a, and their interaction occurs in the PM but not cytoplasm.


Introduction
The high-affinity transport system (HATS) for nitrate uptake in algae and plant is mediated by a two-component NRT2/NAR2 transport system, and some high-affinity NRT2 transporters require the partner protein, NAR2, to function (Quesada et al., 1994;Zhou et al., 2000;Tong et al., 2005;Araki & Hasegawa, 2006;Miller et al., 2007;Cai et al., 2008;Plett et al., 2010;Feng et al., 2011a,b;Yan et al., 2011). Further support for a twocomponent high-affinity nitrate influx system was confirmed using Xenopus oocyte expression analysis. Compared to water-injected Xenopus oocytes, CrNAR2 increased nitrate transport when co-expressed with CrNRT2.1 mRNA (Zhou et al., 2000). Similar results were reported for barley genes, which only transported nitrates when co-injected with both HvNRT2.1 and HvNAR2.3 mRNAs in Xenopus oocytes (Tong et al., 2005). The barley proteins were also investigated, and both were nitrate-inducible and co-localized in the root plasma membrane fraction (Ishikawa et al., 2009). In Arabidopsis, the interactions between AtNRT2.1 and AtNAR2.1 proteins were examined using yeast split ubiquitin and oocyte expression systems, and the results suggested that functional high-affinity nitrate transport may involve an interaction between AtNRT2.1 and AtNAR2.1 (Orsel et al., 2006). However, not all members of the AtNRT2 family require a second gene product for functional nitrate uptake Kotur et al., 2012). All NRT2s, excluding AtNRT2.7, restored growth and b-galactosidase activity in the yeast split ubiquitin system and split yellow fluorescent protein (YFP) fluorescence in Arabidopsis thaliana protoplasts when co-expressed with AtNAR2.1 (Kotur et al., 2012). It was recently reported that OsNAR2.1 is a partner in a two-component nitrate transport system with OsNRT2.1/OsNRT2.2 in rice (Oryza sativa) and OsNRT2.3a in the Xenopus oocyte and yeast-two hybrid system (Feng et al., 2011a,b;Yan et al., 2011).
In barley, both HvNRT2.1 and HvNAR2.3 localized in the plasma membrane (PM) (Wirth et al., 2007), and HvNRT2 and HvNAR2 proteins co-localized in the PM of barley roots (Ishikawa et al., 2009). AtNRT2.1 is mainly localized in the PM of root cortical and epidermal cells, and is involved in the expression of AtNAR2.1 in the PM Wirth et al., 2007). In Arabidopsis protoplasts, NRT2.1 and NAR2.1 polypeptides interact directly at the PM (Yong et al., 2010). Kotur et al. (2012) reported that, with the exception of AtNRT2.7, all remaining NRT2 transporters interacted strongly with AtNAR2.1, mostly at the PM. In rice, Tang et al. (2012) reported that OsNRT2.3a is a PM protein.
A study on the interaction motif in the two-component system using affinity column binding analysis and recombinant proteins was suggestive of a role of the carbon (C)-terminus of HvNRT2.1 in binding to the central region of HvNAR2.3 (Ishikawa et al., 2009). Furthermore, Kawachi et al. (2006) reported that the replacement of Asp105 in the middle region of AtNAR2.1 significantly reduced nitrate uptake and accumulation in Arabidopsis plants.
Although a K(2)K(2)LCY(2)S(3)RxWR(3)D(4)DK motif defining the NAR2 family was originally identified (Tong et al., 2005), the motifs of NAR2.1 proteins that interact with NRT2 proteins for their PM localization and nitrate transporter activity have not been functionally characterized. Here, we used the yeast two-hybrid system, Xenopus oocyte nitrogen-15 ( 15 N) uptake experiments, and the split YFP-labeling system in rice blade protoplasts to examine the interaction between OsNAR2.1 and OsNRT2.3a. To identify the interaction motif of OsNAR2.1, we generated different C-terminal deletions and point mutations in the central region of OsNAR2.1. Green fluorescent protein (GFP) fused to the N-and C-terminus of OsNAR2.1 was used to determine the localization of OsNAR2.1 expression in rice cells.  Fig. S1(a,b). The original OsNAR2.1 in the pBT3-C construct was generated as described previously (Yan et al., 2011). Reverse PCR based on OsNAR2.1 in the pBT3-C construct was used for point mutation construction, and the primers used are listed in Supporting Information Table S1.

Split ubiquitin protein-protein interaction analysis
Protein-protein interactions were examined using a DUAL membrane pairwise interaction kit (Dualsystems Biotech AG, Schlieren, Switzerland) (Yan et al., 2011). Full-length cDNAs of the OsNAR2.1 and OsNAR2.1 mutations were cloned in-frame with the Nub sub-domain of ubiquitin into pBT3-C (LEU2: L, KanR) expression vectors. The primers for subcloning of the C-terminal deletion or point-mutated OsNAR2.1 into pBT3-C OsNAR2.1 are listed in Supporting Information Table S1. The OsNRT2.3a cDNA was cloned in-frame with the Cub sub-domain of ubiquitin in pPR3-N (TRP1: W, AmpR) expression vectors using primers were described previously (Yan et al., 2011). Point mutations of OsNAR2.1 were created at the base of OsNAR2.1 in pBT3-C and all OsNAR2.1 point and deletion mutations in pBT3-C were respectively co-introduced with OsNRT2.3a in pPR3-N into the yeast strain NMY51. Two reporter genes (His and Ade: H and A) allowed the yeast to grow on selective medium (SD-AHLW), after which b-galactosidase (Lac Z) activity assays were performed (Yan et al., 2011).
Functional assay of OsNAR2.1 and OsNRT2.3a in Xenopus oocytes mRNA synthesis of OsNAR2.1 with point mutations and OsNRT2.3a cDNAs, as well as analyses of 15 N-NO 3 À uptake in oocytes were performed as described previously (Yan et al., 2011). OsNAR2.1 and OsNRT2.3a were codon optimized and synthesized by Genescript Company (Nanjing, China). cDNA optimization for Xenopus oocyte expression and subcloning into pT7Ts, oocyte preparation, mRNA injection, and 15 N-NO 3 À uptake assays were conducted as described previously (Feng et al., 2013). Oocytes injected with the genes and water were incubated for 16 h in MBS containing 0.5 mM 15 N-NO 3 À . Primers for gene subcloning into pT7Ts are listed in Supporting Information Table S2.

Constructs for the BiFC assay
OsNRT2.3a and OsNAR2.1 were tagged with the N-and Cterminal halves of EYFP using pSAT vectors (Citovsky et al., 2008). Full-length coding sequence for OsNAR2.1 was fused in-frame with the N-terminal half of EYFP in the PSAT1-nEYFP-C1 vector (nEYFP-OsNAR2.1), which included residues 1-174 of EYFP. The OsNRT2.3a cDNA was fused in-frame to the C-terminal half of EYFP in PSAT1-cEYFP-N1 (OsNRT2.3a-cEYFP), which included residues 175-240 of EYFP. The primers used are listed in Supporting Information Table S3. Plasmid pairs were introduced into rice blade protoplasts using the methods described previously ( and PSAT1-cEYFP-N1 was used as a control for YFP fluorescence assays.

Constructs for transient expression of OsNAR2.1
To examine subcellular localization of OsNAR2.1, we constructed N-terminal GFP fusions (GFP-OsNAR2.1) and C-terminal GFP (OsNAR2.1-GFP) driven by the cauliflower mosaic virus 35S promoter and transfected the derived expression vector into rice protoplast and tobacco epidermis cells.

Fluorescence microscopy
Briefly, 0.2 ml of protoplast suspension of rice blade (c. 2 9 10 cells) was transfected with DNA from 35S:GFP-NAR2.1 and 35S:NAR2.1-GFP constructs (10 lg each). The transformed cells were incubated for 16 to 18 h at room temperature before examination in protoplast medium (R2S + 0.4 M mannitol, Nelson et al., 2007). To mark the PM location, the transfected protoplasts were stained with 5 lg ml À1 of the PM-selective dye marker FM4-64 (Tang et al., 2012). Protoplasts were observed under a 960 objective. GFP expression was visualized by confocal laser scanning microscopy (LSM410; Zeiss AG, Oberkochen, Germany) with a 543 nm helium-neon laser and a 488 nm argon laser, and fluorescent images were captured with a SPOT camera (Tang et al., 2012).
To confirm the location of the 35S:GFP-NAR2.1 and 35S: NAR2.1-GFP fusions in rice protoplasts, we performed transient expression in leaves of tobacco (Nicotiana benthamiana) mediated by Agrobacterium tumefaciens. The combination of the Agrobacterium EHA105 bacterial cells with the indicated constructs above were cultured to an OD 600 of 1.0 and then infiltrated into leaves of tobacco plants. Infiltrated tobacco plants were incubated for another 3-4 d before fluorescence observation (Liu et al., 2012). GFP fluorescence in the cells was analyzed with a 488 nm argon laser using a confocal laser-scanning microscope LSM410 (Zeiss AG).
For BiFC assays, EYFP fluorescence was visualized with a laser wavelength of 525 nm using the confocal microscope described earlier at 24 h after protoplast bombardment of the plasmid pair (Citovsky et al., 2008).

Western blotting
Rice seedlings were grown in IRRI nutrient solution for two months (Yan et al., 2011). The PMs were separated from the other membranes by aqueous two-phase partitioning according to the method of Yan et al. (2011). The microsome (M) fraction and endomembrane (EM) fraction were separated according to the method of Ishikawa et al. (2009). All tissue samples were homogenized and lysed in buffer containing 1% Nonidet P-40 and protease inhibitors. Lysates were cleared by centrifugation and protein concentration was measured spectrophotometrically at A594 nm using Bradford reagent (Sigma-Aldrich, Shanghai, China).
Protein (50 lg) of each sample was boiled in gel loading buffer and resolved on 10% SDS-PAGE gels. Reactive proteins were detected with secondary antibodies including anti-OsNRT2.3a (1:500 dilution, Tang

Subcellular localization of OsNAR2.1
We constitutively expressed the fusion protein OsNAR2.1 and GFP using the pSAT6-EGFP-C1 and pSAT6A-EGFP-N1 expression vector (Tzfira et al., 2005) to investigate the subcellular localization of OsNAR2.1. We transfected OsNAR2.1-GFP and GFP-OsNAR2.1 fusions into rice blade protoplasts under control of the cauliflower mosaic virus 35S promoter, and GFP expression was determined using confocal microscopy. To further explore the localization of OsNAR2.1, we expressed the OsNAR2.1-GFP and GFP-OsNAR2.1 constructs in tobacco epidermis cells (Supporting Information Fig. S3). GFP signaling demonstrated that the fusion protein was transiently expressed in both PM and cytoplasm whether the GFP was fused to the N-or C-terminus of OsNAR2.1 ( Fig. 3a-d, Supporting Information Fig. S3).
To account for the possible effect of GFP on OsNAR2.1 protein localization, we explored the OsNAR2.1 and OsNRT2.3a protein levels in different cellular membranes by Western blotting with specific antibodies. We found that OsNAR2.1 was located in the microsomes (M), which included the endoplasmic reticulum (ER), PM, and EM (mainly including ER) (Fig. 3e), while OsNRT2.3a was located mainly in the PM (Fig. 3e).

Discussion
HATS activity in plants is controlled by a two-component NRT2/ NAR2 transport system, and NAR2 is a partner protein for the NRT2 to take up nitrate (Feng et  ). An Arabidopsis mutant (rnc1) with a mutation (D105N) in the central region of AtNAR2.1 (AtNRT3.1), which was very important for nitrate uptake, was isolated previously (Kawachi et al., 2006). However a complete detailed description of the biochemical activity of this mutation and whether it disrupted the interaction with AtNRT2 partner protein remains unknown. Alignment of NAR sequences among various species showed that this mutation (D105N) occurred in a conserved region of NAR2.1 among different higher plants, and it is also conserved in OsNAR2.1 (D109) (Supporting Information Fig. S1).
In the present study, point mutations were introduced into the conserved regions of OsNAR2.1 to identify the interacting motifs with NRT2 partners (Supporting Information Fig. S1b). We found that two amino acids, R100 and D109, located in the central region of OsNAR2.1 are essential for the functional interaction with OsNRT2.3a for nitrate transport.
Initially, membrane yeast two-hybrid interactions indicated that point mutations W66G, V85F, C88G, K101F, R144G, A150G and G158R did not affect the interaction between OsNAR2.1 and OsNRT2.3a (Fig. 1c,f, Supporting Information Fig. S2). Removing either 7 AAs or 28 AAs from the C-terminal end of OsNAR2.1 protein did not affect its interaction with OsNRT2.3a, indicating that the C-terminus is not important for the interaction (Fig. 1b,e). However, the R100G, R100K or D109N mutant yeast colonies were unable to grow in SD-AHLW medium, suggesting that the two AAs may affect these interactions (Fig. 1a,d, Supporting Information Fig. S2c,d).
Second, to confirm these yeast results, we conducted an 15 N uptake experiment in an oocyte system. Co-injection of OsNAR2.1 with a point mutation (R100G or D109N) and OsNRT2.3a mRNA showed no increase in nitrate transport activity when compared to water-injected controls (Fig. 2). These experiments revealed that R100 and D109 in OsNAR2.1 play important roles in the regulation of HATS activity.
Third, we used transient expression of OsNAR2.1-GFP or GFP-OsNAR2.1 in protoplasts of rice blades and tobacco epidermal cells. Interestingly, we found that OsNAR2.1 was transiently expressed not only in the PM but also in the cytoplasm whether the GFP fusion was on the N-or C-terminus of OsNAR2.1 (Fig. 3,  Supporting Information Fig. S3). Using antibodies and Western blot analysis, we found that the expression of OsNAR2.1 was abundant in both the PM and ER (Fig. 3e), confirming the subcellular localization of OsNAR2.1 in rice protoplast (Fig. 3) and transiently transformed tobacco epidermal cells (Supporting Information Fig. S3). HvNAR2 protein was also examined in both microsomes (M) and PM fractions from barley roots by Ishikawa et al. (2009), although they found that NAR2 was expressed mainly in the PM. Both Ishikawa et al. (2009) and our data based on Western blotting of NAR2.1 indicated that NAR2.1 protein was expressed not only in the PM but also in microsomes (M). Furthermore, OsNAR2.1 was also observed in the EM (mainly ER), which indicated that OsNAR2.1 may remain longer in the ER than HvNAR2.1 during trafficking from the cytosome to the PM.
Because the GFP fusion on the N-and C-terminus of OsNAR2.1 showed a similar expression pattern (Fig. 3, Supporting Information Fig. S3), we suggested that the N-or C-terminus of OsNAR2.1 had no specific sequence for signaling PM location. Based on the GFP expression pattern, we deduced that when we performed BiFC, cYFP-OsNAR2.1 fusion did not damage the original signal for localization of OsNAR2.1. However, as shown in Fig. 4, OsNAR2.1 and OsNRT2.3a co-localized in PM; thus, we explored whether R100 and D109 of OsNAR2.1 are important in targeting OsNAR2.1 and OsNRT2.3a to the PM. As expected, split YFP-labeled OsNRT2.3a and OsNAR2.1 mutants of R100G and D109N did not show any YFP fusion signal in PM (Fig. 4,  Supporting Information Fig. S3). The BiFC results suggested that R100 and D109-dependent OsNAR2.1 binding with OsNRT2.3a might be involved in the targeting of OsNAR2.1 to the PM before the partners interact for nitrate transport.