Rice DUR3 mediates high-affinity urea transport and plays an effective role in improvement of urea acquisition and utilization when expressed in Arabidopsis


  • Wei-Hong Wang,

    1. Department of Plant Nutrition, Key Laboratory of Plant and Soil Interactions of MEoC, College of Resources and Environmental Sciences, China Agricultural University, 100193 Beijing, China
    Search for more papers by this author
  • Barbara Köhler,

    1. Molecular Biology, University of Potsdam, Karl-Liebknecht-Street 24-25, Haus 20, 14476 Potsdam, Germany
    Search for more papers by this author
  • Feng-Qiu Cao,

    1. Department of Plant Nutrition, Key Laboratory of Plant and Soil Interactions of MEoC, College of Resources and Environmental Sciences, China Agricultural University, 100193 Beijing, China
    Search for more papers by this author
  • Guo-Wei Liu,

    1. Department of Plant Nutrition, Key Laboratory of Plant and Soil Interactions of MEoC, College of Resources and Environmental Sciences, China Agricultural University, 100193 Beijing, China
    Search for more papers by this author
  • Yuan-Yong Gong,

    1. Department of Plant Nutrition, Key Laboratory of Plant and Soil Interactions of MEoC, College of Resources and Environmental Sciences, China Agricultural University, 100193 Beijing, China
    Search for more papers by this author
  • Song Sheng,

    1. Department of Plant Nutrition, Key Laboratory of Plant and Soil Interactions of MEoC, College of Resources and Environmental Sciences, China Agricultural University, 100193 Beijing, China
    Search for more papers by this author
  • Qi-Chao Song,

    1. Department of Plant Nutrition, Key Laboratory of Plant and Soil Interactions of MEoC, College of Resources and Environmental Sciences, China Agricultural University, 100193 Beijing, China
    Search for more papers by this author
  • Xiao-Yuan Cheng,

    1. Department of Plant Nutrition, Key Laboratory of Plant and Soil Interactions of MEoC, College of Resources and Environmental Sciences, China Agricultural University, 100193 Beijing, China
    Search for more papers by this author
  • Trevor Garnett,

    1. The Australian Centre for Plant Functional Genomics and the University of Adelaide, PMB1, Glen Osmond, SA 5064, Australia
    Search for more papers by this author
  • Mamoru Okamoto,

    1. The Australian Centre for Plant Functional Genomics and the University of Adelaide, PMB1, Glen Osmond, SA 5064, Australia
    Search for more papers by this author
  • Rui Qin,

    1. College of Life Sciences, South Central University for Nationalities, 430074 Wuhan, China
    Search for more papers by this author
  • Bernd Mueller-Roeber,

    1. Molecular Biology, University of Potsdam, Karl-Liebknecht-Street 24-25, Haus 20, 14476 Potsdam, Germany
    Search for more papers by this author
  • Mark Tester,

    1. The Australian Centre for Plant Functional Genomics and the University of Adelaide, PMB1, Glen Osmond, SA 5064, Australia
    Search for more papers by this author
  • Lai-Hua Liu

    1. Department of Plant Nutrition, Key Laboratory of Plant and Soil Interactions of MEoC, College of Resources and Environmental Sciences, China Agricultural University, 100193 Beijing, China
    Search for more papers by this author

Author for correspondence:
Lai-Hua Liu
Tel: + 86 10 62734463
Email: LL1025@cau.edu.cn


  • Despite the great agricultural and ecological importance of efficient use of urea-containing nitrogen fertilizers by crops, molecular and physiological identities of urea transport in higher plants have been investigated only in Arabidopsis.
  • We performed short-time urea-influx assays which have identified a low-affinity and high-affinity (Km of 7.55 μM) transport system for urea-uptake by rice roots (Oryza sativa).
  • A high-affinity urea transporter OsDUR3 from rice was functionally characterized here for the first time among crops. OsDUR3 encodes an integral membrane-protein with 721 amino acid residues and 15 predicted transmembrane domains. Heterologous expression demonstrated that OsDUR3 restored yeast dur3-mutant growth on urea and facilitated urea import with a Km of c. 10 μM in Xenopus oocytes.
  • Quantitative reverse-transcription polymerase chain reaction (qPCR) analysis revealed upregulation of OsDUR3 in rice roots under nitrogen-deficiency and urea-resupply after nitrogen-starvation. Importantly, overexpression of OsDUR3 complemented the Arabidopsis atdur3-1 mutant, improving growth on low urea and increasing root urea-uptake markedly. Together with its plasma membrane localization detected by green fluorescent protein (GFP)-tagging and with findings that disruption of OsDUR3 by T-DNA reduces rice growth on urea and urea uptake, we suggest that OsDUR3 is an active urea transporter that plays a significant role in effective urea acquisition and utilisation in rice.


Cereal crops are generally inefficient in their uptake and use of soil nitrogen (N) with, for example, only 30–40% of the applied N absorbed by paddy rice (Kronzucker et al., 1998; Raun & Johnson, 1999; Miller & Cramer, 2004). This inefficiency has a significant impact on farm economies and environmental quality. Thus, improvement of crop N-use efficiency has long been recognized as a valuable research topic and a crucial factor contributing to sustainable agriculture and environmental protection.

Although ammonium (NH4+) and nitrate (NO3) represent major soil-N species preferentially used by plants (Jackson & Volk, 1992; Miller & Cramer, 2004), urea can serve as a rapidly available N-source for plant growth (Wang et al., 2008). The degradation of nitrogenous compounds from animals generates great amounts of urea in nature (Wang et al., 2008). In addition, agricultural N-fertilization has led to a tremendous urea input into the biosphere with an annual amount of over 50 million tons accounting for > 50% of the world N-fertilizer consumption (http://www.fertilizer.org/ifa/Home-Page/STATISTICS/Production-and-trade, 2008). However, in natural environments urea accumulation is generally low (< 70 μM in farming soils, Wang et al., 2008), because of the ubiquitous existence of urease, a nickel (Ni)-activated urea-hydrolysing enzyme (Watson et al., 1994). The accessibility of urea to plants is thought to depend mainly on soil microorganism-derived urease activity, which generates NH3/NH4+ that can be subsequently taken up and assimilated by plant cells. However, this process is inevitably associated with the N-loss owing to NH3 volatilization, nitrification and denitrification (Miller & Cramer, 2004). Field experiments showed that the application of urease inhibitors in paddy soils could increase N recovery of rice plants fertilized with urea (Freney et al., 1993), perhaps owing to slow release of NH3 from urea in the soils, thus reducing the rate of N loss. The increase in urea uptake by using urease inhibitors could also be caused by increased absorption of urea directly by roots before its degradation.

The direct absorption of urea as a N source by plants was suggested by several independent physiological observations. Reduction of plant endogenous urease activity, which was caused by Ni depletion in a nutrient solution or using inhibitors or genetic disruption of the urease gene, caused a higher urea accumulation in crop plants (e.g. rice, soybean and potato) when externally provided with urea to either roots or leaves (Stebbins et al., 1991; Gerendás et al., 1998; Witte et al., 2002), suggesting that an increase in urea content in those crops resulted at least in part from the uptake of external urea by plant cells. A pot-soil experiment using [14C]- and [15N]-urea revealed that rice roots did absorb molecular urea, which represented c. 9% of total N uptake from applied urea-N (Safeena et al., 1999). Profiling of amino acids in upper part tissues of Arabidopsis supplied with [15N]urea and 15NH4+ tracers provided evidence that urea and NH4+ shared similar assimilation pathways in the cells (Merigout et al., 2008a), again strengthening the likelihood of a direct use of urea as a N source by the plants. Thus, study of urea transport processes in crops is of agricultural and ecological significance.

To date, molecular identities of urea transporters in higher plants have been studied only in the dicot-model plant Arabidopsis thaliana. The transporter-mediated urea uptake in Arabidopsis roots with a high affinity and low affinity was recently affirmed by quantitative and short-time influx assays of [15N]urea (Kojima et al., 2007). Molecularly, two types of urea-transporting proteins (i.e. DUR3 orthologs and certain MIP members) have been reported in planta (Kojima et al., 2006). Arabidopsis DUR3 (At5g45380), which belongs to the sodium-solute symporter (SSS) superfamily, was characterized as a urea/H+ symporter with a Km of 3–4 μM for urea transport (Liu et al., 2003a; Kojima et al., 2007). The transcriptional upregulation of AtDUR3 in roots under N deficiency and by urea resupply after plant N starvation, as well as the loss of a high-affinity urea transport component observed in the atdur3-mutant lines emphasizes that AtDUR3 may be a major active urea transporter in Arabidopsis (Kojima et al., 2007). Regarding low-affinity urea transporters, a number of isoforms from the PIP, TIP and NIP subfamilies of plant aquaporins were reported to facilitate urea movement with a linear concentration dependency, but their urea transport activity was only evident in heterologous systems (e.g. yeast and frog oocytes) (Gerbeau et al., 1999; Klebl et al., 2003; Liu et al., 2003b). Most recently, the overexpression of AtTIP4;1 driven by the CaMV 35S promoter was reported to enable Arabidopsis leaf cells to absorb more externally supplied [14C]urea, but no distinguishable phenotype of growth and development could be observed in the transgenic lines compared with their wild type (Kim et al., 2008).

We report here for the first time among crop plants the physiological and molecular characterization of urea uptake by OsDUR3 from rice. Kinetic properties of urea influx into rice roots were examined. Heterologous expression of OsDUR3 allowed us to show the functionality of OsDUR3 in urea transport with a high substrate affinity. Moreover, our data gained from protein subcellular localization studies, gene expression analysis and transgenic approaches performed to assess the physiological function of OsDUR3 provide strong evidence that OsDUR3 is a crucial genetic factor for urea transport and plays a significant physiological role in the effective use of low external urea as a N source by rice plants.

Materials and Methods

Plant culture, [15N]urea influx studies and leaf urea determination

Rice (Oryza sativa L.ssp. japonica cv Nipponbare or cv ZH11) seed surface sterilization and germination, and plant growth conditions as well as N-free basic nutrients (plus 1 mM (NH4)2SO4 as N source) used for a normal rice culture (in hydroponics or on agarose-medium) were as described in Cao et al. (2010). For water-logging growth, plants were cultivated in pots (0.19 m diameter, 0.3 m high) filled with 3.5 kg agricultural-soil immersed in water, and the same nutrient solution (1 l) was applied at a tillering- and elongation-stage. Arabidopsis thaliana (Col-0) was grown as described in Liu et al. (2003a). In hydroponics, nutrient solution was refreshed every 2 d. For analyses of rice growth on different N forms, 2 d after germination seedlings were transferred for 8 d to 1 l plastic jars (20 cm high) filled with 80 ml sterile agarose-medium, which contained the N-free basic nutrients with a supply of different N nutrition (no N, 1 or 5 mM urea, 1 mM (NH4)2SO4 or NH4NO3), 1 μM NiSO4, 3% sucrose and 0.4% agarose (Invitrogen).

For the N-dependent gene expression study, plants were cultivated in the normal solution for 15, 16 or 17 d. Except for the control treatment with a continuous supply of 1 mM (NH4)2SO4, plants were starved of N for 1 or 2 d by the addition of 1 mM K2SO4 instead of 1 mM (NH4)2SO4. After 2 d of N starvation, the plants were resupplied with 1 mM (NH4)2SO4 or urea for 3 h. The roots and shoots were harvested, frozen immediately with liquid nitrogen and stored in a −80°C freezer for subsequent total RNA isolation and expression studies.

Root urea-influx assays were conducted with 2-d N-starved plants (see the previous paragraph describing the N-dependent gene expression study). The plants were transferred to 2 l influx assay-solution containing N-free basic nutrients plus [15N]urea (98.65%15N abundance; Novachem PTY Ltd, Melbourne, Australia; for urea ≥ 1 mM, only 50% [15N]urea was applied). After a 8-min exposure of the roots to the assay-solution, the roots were washed three times in 10 mM [14N]urea for 2 min, detached from shoots, dried in an oven at 60°C for 7 d and ground to fine powder. Samples of approx. 3–4 mg were used to determine 15N content using mass spectrometry (Hydra 2020; Sercon, Crewe, UK).

For leaf urea measurement, flag leaves and the third leaves below the flag leaf were harvested from osdur3-dT mutant and its wild-type (ZH11), which were planted in a same soil-pot for 112 d under growth conditions (see the first paragraph in this section describing water-logging growth; Cao et al., 2010). After grinding of leaves with liquid N, 0.1 g power was mixed with 500 μl extraction buffer containing 50 mM HEPES (pH 7.5), 150 mM NaCl, 5 mM EDTA, 0.2 mM 4-(2-aminoethyl) benzenesulfonyl fluoride, 10 mM dithiothreitol and 100 μM phenylphosphorodiamidate (PPD; Invitrogen). Samples were centrifuged by 16 000 g at 4°C for 20 min. 200 μl supernatant was transferred to a fresh tube with 100 μl chloroform, mixed well and incubated at 4°C for 20 min. After centrifugation (16 000 g, 4°C, 20 min), 300 μl color-reagent (Kollöffel & Van Duke, 1975) was added to a 100 μl supernatant sample and incubated at 80°C for 30 min; finally, samples were cooled with ice water for 2 min and their optical density (OD527) was measured using a microplate reader (Infinite 200; Tecan Australia Pty Ltd, Clontarf, Queensland, Australia). Samples of known urea concentration were treated in the same way and used to construct a standard curve.

Yeast functional complementation

Yeast transformation and complementation test were as described in Liu et al. (2003a). All transformants were first selected on SD agar(Oxid)-medium (uracil-deficient yeast nitrogen-base; Difco, Detroit, USA) containing 10 mM NH4+ (i.e. 5 mM (NH4)2SO4) as N source. A single colony from each transformation was picked for growth complementation test on urea as an only N source. The pH of the medium was adjusted with 1 M KOH or HCl.

Uptake experiments in Xenopus laevis oocytes

The OsDUR3 open reading frame (ORF) was cloned into the vector pOO2 (Liu et al., 2003a) via BamHI and (vector-)BglII (compatible with BamHI) restriction sites, yielding the plasmid pOO2-OsDUR3. Capped cRNA was transcribed in vitro from pOO2-OsDUR3 (Liu et al., 2003a). Oocytes were removed from adult female frogs by surgery and dissected manually. The oocytes (Dumont stage V or VI) were defolliculated using 10 mg ml−1 collagenase (Sigma) for 1 h and injected with 50 nl of cRNA (1 μg μl−1). After injection the oocytes were kept for 3–4 d at 20°C in ND96 solution (Liu et al., 2003a). Data were collected from several batches of oocytes from different frogs as indicated in Fig. 3.

Radiotracer uptake studies in oocytes were performed according to Liu et al. (2003a). Radioactivity was monitored using a liquid scintillation counter (LS6500; Beckman Coulter Krefeld, Germany). Standard bath solutions contained 100 mM choline-Cl, 2 mM CaCl2, 2 mM MgCl2 and 4 mM Tris, and the pH adjusted to 5, 6, 7 or 8 with MES; 100 μM thiourea and 1 mM putrescine were added for urea transport competition assays.

Total RNA isolation and gene expression analysis by qPCR

Total RNA of different rice tissues (e.g. roots, shoots) was isolated using TRIzol method (Invitrogen). RNA from germinating seeds was extracted (Liu et al., 2003a). The RNA samples were pretreated with 1 unit DNaseI (Invitrogen) for 30 min before performing reverse transcription in vitro. One microgram of RNA was used to synthesize cDNA by reverse transcriptase using SuperScript III Reverse Transcriptase, according to the manufacturer’s protocol (Invitrogen). Quantitative PCR (qPCR) was conducted in a 20 μl volume (containing 2 μl of 1 : 10 diluted original cDNAs, 200 nM of each gene-specific primer, and iQ SYBR Green Supermix; Bio-Rad) using a Bio-Rad iCycler. The following primers were used for qPCR or semi-quantitative PCR (Liu et al., 2003b): OsDUR3, 5′-CCTTGGCTACTTCACGCTGT-3′ and 5′-TGCATCTCCGTCTCGTGTAG-3′; OsGS1.2 (Ishiyama et al., 2004), 5′-AAAGGCGTTCGGCCGCGACATCGTGGAC-3′ and 5′-CACTTGGTCAGCAGCGGCGATGCCAACT-3′; Osarginase, 5′-ACAGAAGAAGGCAAAGAACTCA-3′ and 5′-GACCAATGGCCGAAGAGGAT-3′; OsUrease, 5′-ATTACTCGGACATGGCAAAC-3′ and 5′-AAAGCCATTCACTATAGCAGGA-3′; OsGAPDH (Kim et al., 2003), 5′-GGGCTGCTAGCTTCAACATC-3′, 5′-TTGATTGCAGCCTTGATCTG-3′; Ostubulin (Kim et al., 2003), 5′-TACCGTGCCCTTACTGTTCC-3′ and 5′-CGGTGGAATGTCACAGACAC-3′; OsAMT1.2 (Kumar et al., 2003), 5′-TGAAGCACATGCCGCAGAC-3′ and 5′-CGGTGAGGAAGGCGGAGTA-3′; Os03g48930 gene, 5′-GGATGTTGGTGCCTCTTTGT-3′ and 5′-GCATACTCCCTAGCACCCT-3′; Os03g48940 gene, 5′-CTCAGTGGCAGCACACATCT-3′ and 5′-TCTGTCGCATAGGATTGCAG-3′; OsACTIN1, 5′-ATGGCTGACGCCGAGGATATCCA-3′ and 5′-TTAGAAGCATTTCCTGTGCACAATG-3′. The correctness of the PCR amplicons was verified by sequencing.

Protein localization in Arabidopsis protoplasts

The OsDUR3 ORF without stop codon was amplified by Pfu DNA polymerase (Invitrogen) using primers (5′-AGagatctATGGCGAGCGGCGTGTGCCC-3′, 5′-ATagatctgGGAGTGCATCATCTGATTATTATT-3′), and cloned into pCF203 (Liu et al., 2003b) using the BamHI site, yielding an OsDUR3:GFP construct. AtDUR3:GFP and AtTIP4;1:GFP fusions were generated previously (Liu et al., 2003b). The GFP (green fluorescent protein) fusion constructs were introduced into Arabidopsis protoplasts (Col-0) prepared from a 7-d-old cell culture (Liu et al., 2003b). Protoplasts were analysed for fluorescence 24–48 h after transformation using a confocal laser scanning microscope (Leica DM IRBE with Leica TCS SPII laser scanning unit; Leica, Wetzlar, Germany) using a ×10/×63 water objective lens. Fluorescence excitation was achieved with an Ar/HeNe laser at 488 nm. Images were arranged using Leica software/Adobe Photoshop (Adobe Systems, Mountain View, CA, USA).

Generation of OsDUR3-transformed Arabidopsis lines, growth phenotyping and [15N]urea root uptake

The OsDUR3 ORF was amplified by Pfu using two primers containing BglII and SalI site, respectively (5′-AAGagatctATGGCGAGCGGCGTGTGCCCTC-3′, 5′-TGGgtcgacTCAGGAGTGCATCATCTGATT-3′), and cloned into the nonGFP pCF203 using BglII/SalI ends, to obtain a construct termed pCF203(-):OsDUR3. Arabidopsis Col-0 and atdur3-1 plants (Kojima et al., 2007) were transformed by dipping inflorescences into a cell suspension (OD600 = 0.6) of Agrobacterium GV3101 harboring pCF203(-):OsDUR3. Harvested seeds were selected on kanamycin (50 mg l−1)-containing 1/50 B5 agar-medium to obtain transformants. Several independent homozygous OsDUR3-transformed lines were generated in T2 or T3 generation for experimental use.

For the growth complementation test, surface-sterilized seeds were grown for 16 d on the 1/50 B5 agar-medium containing different concentrations of urea or NH4NO3 with a supplementation of 1 μM NiSO4. To detect OsDUR3 (over-) expression in Arabidopsis, reverse-transcription polymerase chain reaction (RT-PCR) was performed on total RNA isolated from Col-0, atdur3-1 and transgenic lines (Liu et al., 2003a). Primers used for the RT-PCR were: OsDUR3, 5′-GTCGTCTTCGTCTTCCTCGTC-3′ and 5′-GCTCATCCAGTAGCCGTTGTC)-3′; AtACT2, 5′-TCCAAGCTGTTCTCTCCTTG-3′ and 5′-AGACGGAGGATGGCATGAG-3′ (Liu et al., 2003a).

The root urea uptake study was performed with Col-0 and OsDUR3-overexpession lines grown hydroponically for 3 wk. The protocol is as described earlier in ‘[15N]urea influx studies’.

Isolation and analysis of rice osdur3-dT mutant

A T-DNA insertion line for OsDUR3 (acc. no. 03Z11CL14; background ZH11 (Oryza sativa L. ssp. japonica)) was purchased from the National Center of Plant Gene Research of China (http://rmd.ncpgr.cn/). Homozygous mutant was identified by PCR using the primers: a T-DNA left border primer, 5′-AATCCAGATCCCCCGAATTA-3′ and two gene-specific primers, 5′-ATGATGCAGATGTGGAGAGGC-3′, 5′-CTCGCCATCGCCAGCTTCAAT-3′.

To assess T-DNA numbers and locations in the genome of the mutant, thermal asymmetric interlaced-PCR (TAIL-PCR) was conducted (Liu & Chen, 2007). Primers used for the PCR were: (LB1) 5′-AGGGAATTAGGGTTCCTATAG-3′, (LB2) 5′-CATGTGTTGAGCATATAAGAAA-3′, (LB3) 5′-ATTCCTAAAACCAAAATCCAGT-3′, (AD1) 5′-NTCGASTWTSGWGTT-3′, (AD2) 5′-NGTCGASWGANAWGAA3′ and (AD3) 5′-WGTGNAGWANCANAGA-3′. The resulting PCR products were gel-purified, directly sequenced and blast-analysed against the rice database. Two T-DNA insertions were detected in the 03Z11CL14 mutant, one in OsDUR3 gene and another in an intergenic region (Fig. 8a).


Two distinct transport processes for urea are determined in rice roots

The kinetic features of root urea absorption by any crop plant have hitherto not been characterized. We thus applied stable isotope-labeled [15N]urea to examine urea influx into root cells of 3-wk-old and 2-d N-starved rice plants (O. sativa, japonica) at pH 5.8 (for rice optimal growth). Because NH4+ was undetectable in the urea influx assay solution under our experimental conditions (data not shown), similar to a situation reported by Merigout et al. (2008b), we proposed that in the assay solution there would be no urea degradation into NH4+ by, for example, urease. Thus, the accumulation of 15N analysed in the plants can be regarded as being derived from external [15N] urea absorbed by the roots. As urea concentration in natural and farming soils was reported in micromolar ranges (Wang et al., 2008), urea uptake by the roots was measured first at low urea supply (50 μM or 200 μM) in the absence or presence of the protonophore CCCP (carbonyl cyanide m-chlorophenylhydrazone), which is commonly used to dissipate the proton-motive force created by (plasma) membrane-associated proton-ATPases. The application of 50 μM CCCP significantly decreased [15N]-tracer or [15N]-urea uptake in the roots by c. 30% (Fig. 1a), indicating that at least one factor for root urea absorption at micromolar concentrations is active or depends on metabolic energy, which is generated by the proton gradient across the plasma membrane (PM) of rice root cells.

Figure 1.

Physiological characterization of urea influx into rice roots. (a) Reduction of root urea-uptake by the protonophore carbonyl cyanide m-chlorophenylhydrazone (CCCP). Plants with 15 d normal growth and 2 d N-starved plants were used in influx assays (see the Materials and Methods section). Roots were supplied with 50 or 200 μM [15N]urea in the absence or presence of 50 μM CCCP for 20 min. In the control (without CCCP), an equal volume of dimethyl sulfoxide (DMSO) for solubilizing CCCP was added. DW, dry weight. Data are means ± SE, = 3 (biological replicates), *Statistically significant difference (< 0.05 by Student’s t-test). Open bars, without CCCP; tinted bars, CCCP. (b,c) Concentration-dependent root urea-influx. Roots were exposed to [15N]urea assay-solution for 8 min. A saturable or linear concentration-dependent urea uptake by the roots was identified at a urea supply of 5–40 μM (b) and of 40–1000 μM (c). Data are means ± SE, = 4. 15N incorporated in the roots was analysed by mass spectrometry and was converted into urea taken up by the roots.

Time- and concentration-dependent uptake experiments were further conducted to characterize urea transport kinetics of rice. After a 30-min incubation of roots in the assay-solution containing 50 μM [15N]urea, 15N could be detected in shoots, which accounted for c. 3% of urea taken up in the roots (data not shown), suggesting a long-distance transport of 15N in the form of either urea or its assimilation product(s) from roots to shoots. However, after a 15-min exposure of the root to the urea solution, no 15N concentration exceeding its natural abundance was measured in the shoots (data not shown). To minimize the effect of a depletion of root-absorbed urea pool through long-distance translocation and/or urea catabolism on urea uptake, we examined the concentration-dependent urea influx after incubation of roots in [15N]urea solution for only 8 min. In a concentration range of 5–1000 μM, urea influx into the roots was found to follow at least two distinguishable transport pathways. At lower concentrations (5–40 μM), the influx was saturable and resembled mostly Michaelis–Menten kinetics, allowing calculation of an affinity constant (Km) of 7.55 μM for urea import into root cells (Fig. 1b). This result strongly suggests the existence of an active transport process for urea in the rice roots, as shown by the CCCP-inhibited influx at low external urea (Fig. 1a). A linear concentration dependency for root urea influx was determined at higher concentrations that is, 40–1000 μM (Fig. 1c).

An active urea transporter DUR3 ortholog is identified in rice

To understand molecular basis of urea permeation of rice, sequences of so far well-characterized urea-transporting proteins, including mammalian and bacterial urea transporters (UTs, UreI from Helicobacter spp., Yut from Yersinia spp.) and Arabidopsis urea/H+ symporter AtDUR3 (Wang et al., 2008) were used as templates to search for homologs in rice databases (using a blast e-value cut-off of 1e-5). Only one gene at a rice locus_Os10g42960 was identified as an ortholog of AtDUR3. We thus termed this gene OsDUR3. No other gene(s) similar to UTs or UreI was uncovered in the rice genome.

The deduced OsDUR3 shares 74.9% amino acid sequence identity with AtDUR3, contains 721 amino acid residues and is predicted as an integral membrane-protein exhibiting 15 transmembrane-spanning domains (TMSD) with the N- and C-terminus outside and inside cytoplasm, respectively (see the Supporting Information, Fig. S1a). The predicted membrane topology of OsDUR3 is similar to that of AtDUR3 (14 TMSDs). The TMSDs 1–6 and 7–12 seem to be well conserved between the two proteins, but the C-terminal parts appear to be variable in their length (Fig. S1a). The coding sequence of OsDUR3 is derived from three exons, and its intron–exon structure considerably differs from AtDUR3 (Fig. S1b).

We previously isolated and sequenced a rice cDNA of 2478 bp (GenBank no. AY463691), which exhibits a poly-A stretch of 11 adenosine residues. The cDNA possesses a long open reading frame (ORF, 2166 bp) with an ATG start-codon at position 103 bp and a termination-codon TGA at position 2268 bp. This ORF is identical to the coding sequence resulting from the three exons of OsDUR3 as predicted in ‘Aramemnon’ (http://aramemnon.botanik.uni-koeln.de), suggesting that the 2478-bp cDNA harbors the complete OsDUR3 coding region.

OsDUR3 functionally complements a yeast urea uptake-deficient mutant

To examine the molecular function of OsDUR3 in urea transport, the 2166-bp ORF was cloned into a yeast-expression vector pHXT426, and expressed in a yeast mutant YNVW1 defective for urea uptake (< 5 mM. Liu et al., 2003a). Fig. 2 shows that OsDUR3 enabled growth of the YNWV1 on 2 mM urea comparable to that of the wild-type positive control strain 23346c, whereas no growth was observed for YNVW1 transformed with the empty vector, suggesting that OsDUR3 mediates urea movement across the yeast PM. The growth rate of yeast transformants harboring the OsDUR3 cDNA decreased with increasing medium-pH (Fig. 2), indicating that OsDUR3-facilitated urea uptake occurred preferentially at lower external pH, similar to observations made for AtDUR3 and PiDUR3 when expressed in the YNVW1 (Liu et al., 2003a; Morel et al., 2008).

Figure 2.

Functional complementation of yeast mutant YNVW1 by heterologous expression of OsDUR3. The yeast strains 23346c (Δura3) and YNVW1 (Δdur3, Δura3) transformed with an expression-vector pHXT426 alone or harboring OsDUR3-ORF were grown first on SD (yeast N base without ammonium sulfate and without amino acids) medium containing 10 mM NH4+; a single colony was picked, suspended in 30 μl water, serially diluted and spotted onto the SD medium containing 2 mM urea as only N source at different pH. YNVW1 did not grow on < 5 mM urea (Liu et al., 2003a). Photographs were taken after yeast growth on urea medium for 5 d.

OsDUR3 mediates a high-affinity urea import in Xenopus oocytes

The transport nature of OsDUR3 was characterized in X. laevis oocytes, which are often used to investigate protein-facilitated urea transport processes (Leung et al., 2000; Liu et al., 2003a,b). [14C]Urea uptake by OsDUR3-expressing oocytes confirmed the functionality of the transporter. OsDUR3 facilitated high-affinity import of urea across the oocyte PM, with a maximal uptake rate of c. 12 pmol oocyte−1 h−1 (Fig. 3a). Accumulation of [14C]urea by OsDUR3 saturated at c. 80 μM and the concentration of urea permitting a half-maximal uptake was (Km) c. 10 μM (Fig. 3a).

Figure 3.

Biochemical characterization of urea transport by OsDUR3 in Xenopus laevis oocytes. (a) Concentration-dependent [14C]urea uptake was saturable at c. 80 μM external urea and displayed a Michaelis–Menten kinetics with a Km of 10.32 ± 1.53 (SE) μM, and a Vmax of 12.79 ± 0.43 (SE) pmol oocyte−1 h−1. Standard choline-Cl buffer solution at pH 5.0 was applied (also for (c) and (d), see the Materials and Methods section). Means ± SE (= 9 measurements from three batches) are plotted. Values are differences of [14C]urea of injected and noninjected oocytes (also for (b) and (c)). (b) pH-dependence of urea uptake. Mean values ± SE (= 10 measurements from four batches) are plotted. (c) Competition uptake assay in the absence and presence of the urea analog thiourea (Tu) at 100 μM. Mean values (± SE) of five measurements (= 5) from two batches are plotted. (d) Competition uptake assay in the absence (control) or presence of the polyamine putrescine (Pu) at 1 mM. Means ± SE (= 4 measurements) derived from two batches are shown.

Increasing the pH of the medium significantly reduced urea uptake (Fig. 3b), in accordance with the model for urea uptake by OsDUR3 as shown for AtDUR3 in oocytes (Liu et al., 2003a). However, the pH-dependence of urea transport by OsDUR3 differs from that of AtDUR3 as urea uptake was strongly reduced at pH 6 (Fig. 3b). Moreover, thiourea, a structural urea analog, is found to be a competitive inhibitor for urea uptake by OsDUR3 in the oocytes. The addition of 100 μM thiourea caused a decline in urea accumulation by over 50% compared with measurements in the absence of thiourea (Fig. 3c). Recently, Uemura et al. (2007) reported that yeast ScDUR3 transported not only urea but also polyamines (putrescine, Km = 479 μM; spermidine, Km = 21.2 μM). However, substrate competition assays did not yield any evidence that such polyamines affected urea permeation via OsDUR3. The addition of 1 mM putrescine had no effect on urea import (Fig. 3d).

OsDUR3 is localized to the plasma membrane

To clarify a possible physiological role of OsDUR3 in plant urea uptake, an expression construct of OsDUR3 fused C-terminally to GFP under the control of a CaMV 35S-promoter was transiently expressed in Arabidopsis protoplasts. Protoplasts transformed with recombinant plasmids encoding GFP-tagged AtTIP4;1 and AtDUR3 served as references, which were shown to be mainly localized in the tonoplast and PM, respectively (Liu et al., 2003b; Kojima et al., 2007). The green-fluorescence signal derived from OsDUR3:GFP expression was restricted to a thin ring comparable to the size of the protoplasts most similar to the signal resulted from AtDUR3:GFP fusion-protein, whereas fluorescence of GFP fused to the AtTIP4;1 was confined to border areas of the protoplasts and to some internal structures (Fig. 4a). After plasmolysis of the protoplasts by osmotic shock, fluorescence signals of OsDUR3:GFP and AtDUR3:GFP were observed in cellular material that most probably arose from the ruptured PM, whereas the TIP4;1:GFP signal was present in the tonoplast (Fig. 4b). Thus, GFP-labeling revealed a PM-targeted localization of OsDUR3 consistent with its physiological role in urea transport into the plant cells.

Figure 4.

Subcellular localization of OsDUR3 protein fused with green fluorescent protein (GFP). Transmission image (left column), fluorescence image (middle column) and merged image (right column) are shown. (a) GFP fluorescence from Arabidopsis protoplasts transformed with OsDUR3:GFP, AtDUR3:GFP or AtTIP4;1:GFP fusion constructs. (b) GFP fluorescence was visualized after disruption of the protoplast plasma membrane by osmotic shock using 100 mM Na2HPO4 and 10 mM EDTA, pH 6.5.

OsDUR3 mRNA-abundance is regulated by varied N regimes and occurs differently in tissues

As urea occurs as a N source available for plants in soils and is an essential intermediate in plant N metabolism (e.g. produced during arginine degradation by arginase) (Wang et al., 2008), N-dependent gene expression was analysed by quantitative RT-PCR (qPCR) in 3-wk-old rice. OsDUR3 was expressed in roots and shoots of plants grown under different N treatments (Fig. 5a). The amount of OsDUR3 mRNA in the roots was over twofold higher than that in the shoots (Fig. 5a), indicating a possible role for OsDUR3 in urea uptake in rice roots. OsDUR3 transcripts remained at a relatively low level under normal growth conditions with NH4+ -provision, but the 2-d N depletion greatly stimulated OsDUR3 expression (Fig. 5a). Interestingly, a 3-h N resupply with 1 mM urea to the N-starved rice caused a marked increase in the abundance of OsDUR3 mRNA in the roots compared with that observed under N starvation, while NH4+ resupply did not obviously affect OsDUR3 expression (Fig. 5a). This suggests that OsDUR3 transcription is likely upregulated by the substrate of its encoded transporter. To corroborate this N-dependent transcriptional regulation of OsDUR3, the expression of a well-studied N-regulated gene OsGS1.2 was examined as a reference. Upon continuous supply of NH4+ or 3-h NH4+-resupply after N starvation, OsGS1.2 mRNA accumulated strongly in the roots (Fig. 5b), most similar to that reported by Ishiyama et al. (2004). Resupply of urea also elevated OsGLN1;2 transcript abundance in roots, possibly because of its metabolic conversion into NH4+ (Fig. 5b).

Figure 5.

Transcriptional upregulation of OsDUR3 by nitrogen (N) deficiency and urea in rice roots and shoots. Relative mRNA accumulation of (a) OsDUR3 and (b) OsGS1.2 (AB180688, a N-related gene as a reference. Ishiyama et al., 2004) was quantified by quantitative reverse-transcription polymerase chain reaction (qPCR; see the Materials and Methods section). Rice growth and N treatments are described in the Materials and Methods section. Gene-specific primers designed from the third exon of OsDUR3 were used. Means of three biological replicates and two technical repeats ± SE (= 6) are shown. Different letters above the bars indicate statistically significant differences (< 0.05 by Student’s t-test). Rice tubulin (NM_001051100. Kim et al., 2003) was used as an internal reference gene. The mRNA level of OsDUR3 or OsGS1.2 from 2 d N-starved rice leaves was set to 1 for the calculation of relative gene expression.

Moreover, the fluctuation of OsDUR3 expression was monitored in germinating seeds and in different tissues, where N remobilization processes are suggested to be prominent (Limami et al., 2002). Between days 1 and 4 of the germination, OsDUR3 transcripts were detected at a similar low level, whereas a dramatic increase in OsDUR3 mRNA was observed at day 6 after germination (Fig. 6a). A similar expression pattern was also measured for the ammonium transporter OsAMT1.2 and for Osarginase (Kumar et al., 2003; Cao et al., 2010) (Fig. 6a). By contrast, the mRNA abundance of OsUrease appeared not to be obviously changed during a 6-d germination (Fig. 6a), in agreement with observations that the activity of this enzyme remained stable in germinating seeds (Cao et al., 2010). Thus, the transcriptional upregulation and co-regulation of rice DUR3, AMT1.2 and arginase at a later germination stage may reflect the plant response to N starvation and a requirement of the plant for the redistribution of N, as urea and NH4+, from source sites to growing sink organs during seedling development. Furthermore, OsDUR3 expression was relatively high in old leaves, but low in young flag leaves and unopened flowers (Fig. 6b). Such tissue-specific expression patterns were also found for Osarginase (Fig. 6b). The expression of OsUrease in different tissues seemed to be constitutive (Fig. 6b). Thus, the concomitant expression of DUR3 and arginase in old leaves suggests a possible physiological function for OsDUR3 in the remobilization of urea-N from senescing organs to a growing sink part whenever required.

Figure 6.

Expression of OsDUR3 and other nitrogen (N)-related genes in germinating seedlings and different tissues. Quantitative reverse-transcription polymerase chain reaction (qPCR) was performed on total RNA from rice seedlings after germination over (a) a 6-d period or (b) different tissues. Seed germination and plant growth for the tissue-specific gene expression study are described in the Materials and Methods section. FL, flag leaf emerged after 6 d but not fully expanded; OL, old leaf harvested from the third leaves below the flag leaf; Fl, flowers sampled at the time FL was harvested. Gene-specific primers designed from the third exon of OsDUR3 were used. The N-related genes are AMT1.2 (AF289478, ammonium transporter), OsUrease (J023004J18) and Osarginase (NM_001058548). A housekeeping gene GAPDH (Kim et al., 2003. GQ848032) served as an internal reference. Means of two biological replicates and three technical repeats ± SE (= 6) are presented. For the calculation of relative gene expression, mRNA levels of the tested genes in 1-d germinating seedlings or flag leaf were set to 1.

Overexpression of OsDUR3 improves Arabidopsis growth on lower urea and enhances root urea acquisition

Molecular and physiological functions of OsDUR3 in urea absorption and utilization for plant growth were investigated in OsDUR3-transformed homozygous Arabidopsis lines. Plants were grown for 16-d on sterile agar-medium without N supply, or supplemented with urea or NH4NO3 at different concentrations. OsDUR3 transcripts were clearly shown to occur in roots of two representative OsDUR3-transformed wild type (WT) lines (WT+OsDUR3-6 and -7 ) and atdur3-1 mutants (atdur3-1 + OsDUR3-1 and -2), but not in untransformed WT and atdur3-1 plants (Fig. 7a). A complementation assay demonstrated improved growth of OsDUR3 transformants when grown on urea as N source, particularly at low urea (≤1 mM), whereas both WT and atdur3-1 plants grew slowly and developed symptoms of N deficiency (Fig. 7a), most similar to the report by Kojima et al. (2007). At a higher urea concentrations (e.g. 3 mM), all plants grew normally and no apparent growth differences were observed between the different lines (Fig. 7b). Expression of OsDUR3 did not affect the utilization of NH4NO3 as N source by Arabidopsis (Fig. 7b). Other independent transgenic homozygous lines (five and three lines for WT + OsDUR3 and atdur3-1 + OsDUR3, respectively) also displayed similar phenotypes when grown with urea (data not shown).

Figure 7.

Expression of OsDUR3 in Arabidopsis improves plant growth on urea and enhances urea uptake from the medium. OsDUR3-ORF under the control of the CaMV 35S-promoter was transformed into Arabidopsis WT (Col-0) or DUR3-defected mutant atdur3-1 (Kojima et al., 2007). Two representative independent transgenic lines of WT+OsDUR3 or atdur3-1 + OsDUR3 were used. (a) Detection of OsDUR3 expression in the roots of different lines. Semiquantitative reverse-transcription polymerase chain reaction was performed on total RNA from the roots of 3-wk-old plants. Primers for amplification of OsDUR3-ORF were applied. The constitutively expressed Arabidopsis ACTIN2 (AtACT2) served as reference. (b) Growth phenotyping of OsDUR3-transformed lines on different nitrogen (N) sources. Plants were grown for 16 d on nutrient agar-medium containing urea or ammonium nitrate (AN). (c) Enhancement of urea influx into roots of OsDUR3-overexpressing plants. [15N]urea uptake was determined in roots of 3 wk-old wild type (WT) and OsDUR3-overexpression plants at 100 μM [15N]urea root-supply for 20 min. 15N measurement and its conversion in urea is described as in Fig. 1. DW, dry weight. Data represent means ± SE, = 4 (biological replicates, four plants in each), and different letters above the bars indicate statistically significant differences (< 0.05 by Student’s t-test).

When [15N]urea was supplied at 100 μM to roots of 3-wk-old Arabidopsis, two independent transgenic lines accumulated c. sixfold more 15N isotopic tracer than the WT (Fig. 7c), clearly demonstrating that the constitutive (35S-promoter-drived) expression of OsDUR3 greatly enhances urea acquisition across the PM of root cells.

Disruption of OsDUR3 causes growth retardation and less urea uptake of rice plants on urea as the sole N source

To examine the physiological significance of OsDUR3 in rice plant per se, a homozygous T-DNA insertion line (ZH11 background, O. sativa L. ssp. Japonica; see the Materials and Methods section) was isolated. Because only one osdur3-mutant line was obtained from the database, it is necessary to examine T-DNA numbers and their exactly integrated positions as well as the expression of T-DNA inserted or flanking genes in the mutant. Using thermal asymmetric interlaced-PCR we detected two T-DNA insertions in the mutant. One T-DNA was integrated in the OsDUR3 promoter-region (locus Os10g42960) 410 bp upstream of the putative translation start site (Fig. 8a), and another occurred in a repeat-region flanked by two adjacent genes, which encode the putative α-subunit D of 20S proteasome (Os03g48930) and a chloride channel (OsCHL-D; Os03g48940), respectively (Fig. 8a). We termed this rice line osdur3-dT. Gene expression test revealed a quite low level of OsDUR3 mRNA in the osdur3-dT relative to that in the WT (Fig. 8b), indicating a strong OsDUR3 knockdown owing to the T-DNA insertion.

Figure 8.

Identification of rice osdur3-dT mutant line. (a) Schematic representation of T-DNA locations in OsDUR3 and in another genomic-position detected by a thermal asymmetric interlaced-polymerase chain reaction (TAIL-PCR) (see the Materials and Methods section). A T-DNA (upper part) was inserted in OsDUR3-promoter region. The second T-DNA (lower part) was in a repeat region between two adjacent genes (in a 680 bp-distance), one coding for a 20S proteasome α-subunit D (Os03g48930) and another for a putative chloride channel (OsCHL-D, Os03g48940). We named this mutant osdur3-dT (acc. no. 03Z11CL14; in the background ZH11 (WT)). (b) A strong reduction of OsDUR3 expression in the osdur3-dT. Quantitative RT-PCR (qPCR) was performed on total RNA from wild type (WT) and osdur3-dT plants (19 d normal growth and 2 d nitrogen (N)-starved; see the Material and Methods section). Means indicate relative mRNA-accumulation ± SE, = 3 biological repeats. Tubulin (NM_001051100) expression served as an internal control. (c) Expression test of the T-DNA-flanking genes in the osdur3-dT. Reverse-transcription polymerase chain reaction (RT-PCR) (in 25 or 30 cycles) was conducted on total RNA from plants as described in (b). Gene-specific primers designed from two flanking genes (Os03g48930, Os03g48940) were used, allowing amplification of 249- and 211-bp PCR product, respectively. OsACTIN1 (AK100267) served as a control showing that an equal amount of cDNA was used in all PCR.

When grown for 8-d on sterile-medium with even 5 mM urea as N source, osdur3-dT showed obvious growth inhibition characterized by the reduction in plant size/height, slight yellowing of shoots and biomass accumulation of both, roots and shoots compared with the WT (Fig. 9a,b). With NH4NO3 as N source osdur3-dT and WT plants grew similarly well (Fig. 9a,b). These observations suggest that the suppressed growth of osdur3-dT in response to urea (but not NH4NO3) most likely results from the OsDUR3 knockdown, rather than from a physical interruption of the genomic repeat-sequence by the second T-DNA insertion (Fig. 9a), although at a higher urea supply we could not rule out the possibility that a disorder of an internal transport process, for example, N translocation (in the form of asparagine and/or glutamine) from roots to shoots, might become an important factor affecting urea use by rice (Cao et al., 2010). The expression of the two flanking genes was similar in the mutant and WT (Fig. 9c). Finally, the ability of the osdur3-dT to absorb urea from the medium was assessed by a [15N]urea root-influx study in rice grown for 19-d under normal N nutrition and then starved for N for 2 d. When urea was supplied for 8 min at 75 μM or 1000 μM, urea uptake by mutant roots was 20–23% lower than in WT (Fig. 9c), strongly supporting the model that OsDUR3 acts as a transporter significantly contributing to urea acquisition in a range of μM to low mM urea.

Figure 9.

Reduction of growth rate and urea uptake as well as increase in old-leaf urea content observed in the osdur3-dT mutant. (a) Growth test of the osdur3-dT and wild type (WT) (ZH11) on urea and ammonium nitrate (AN). After a 2-d germination, seedlings removed from the rest of the seeds (to limit seed nitrogen (N) provision) were transferred to sterile nutrient-medium (pH 5.8, 0.4% agarose) supplied with 1 μM NiSO4, 3% sucrose, and urea or AN as N source for 8 d growth. (b) Biomass reduction of osdur3-dT grown with urea as N source. Plants from (a) were harvested for biomass quantification. Data represent means ± SE (= 4 replicates, four plants in each). (c) Reduction of urea uptake by roots of the osdur3-dT mutant. Plants were cultivated and short-time (8 min) influx studies using [15N]urea (at 75 or1000 μM) were conducted (as described in Fig. 1). DW, dry weight. Means ± SE (= 3) are shown. (d) Increase in urea content in old leaves of osdur3-dT mutant. Plant growth and urea measurement are described in the Materials and Methods section. All leaf samples were harvested on the same day. WT, wild-type plant; Mu, mutant plant; FL, flag leaves emerged after 5–6 d; OL, old leaves harvested from the third leaf below the flag leaf. FW, fresh weight. Means of 3–4 biological replicates ± SE (= 3–4) are presented. Different letters above the bars indicate statistically significant differences (< 0.05, by Student’s t-test).

Analysis of leaf urea content revealed that old leaves contained significantly more urea than flag leaves in general (Fig. 9d). Further, the accumulation of urea in the old leaves of osdur3-dT mutant was c. 40% greater than that of WT (Fig. 9d). These results lead to a strong suggestion of a requirement of OsDUR3 for the mobilization of internal urea from senescing tissues to young parts.


Rice roots possess a high- and low-affinity transport system for urea uptake

Our work has provided for the first time kinetic evidence for crop urea acquisition. A concentration-dependent influx-study using 15N-tracer showed that rice is equipped with a high-affinity and a low-affinity transport system for urea acquisition (Fig. 1). In a range of 5–40 μM urea, a Michaelis–Menten type transport process with a Km of c. 8 μM for urea was evident, whereas a linear concentration dependence for root urea-influx was determined in the concentration range of 40–1000 μM (Fig. 1b,c). The kinetics of urea uptake in rice is comparable to that observed in Arabidopsis, where a Km of 4 μM for the high-affinity urea-uptake component was measured (Kojima et al., 2007).

Real-time monitoring of 13N-translocation in rice using positron-emitting techniques demonstrated that the process starting from ammonium uptake to its incorporation into glutamine and thereafter the movement of glutamine reaching the junction between root and shoot could be accomplished within 4 min (Kiyomiya et al., 2001). However, because long-distance translocation of the imported (urea-)15N from the roots to shoots was not measurable in our influx experiments (data not shown), we assume that most urea absorbed by the roots should remain in the intact form in root cells. Thus, urea-influx kinetics determined here might reflect native or physiological processes of urea transport by intact rice roots, leading to the suggestion that the existence of both high- and low-affinity urea-uptake systems operating at different urea concentrations should make it possible for rice to directly capture and use urea as a N source once available in growth environments. At the molecular level, in this work we have functionally characterized OsDUR3, which most represents likely a critical component for high-affinity urea transport in rice (see later). Regarding the linear transport process, although its molecular identity in rice is not established, certain member(s) of some gene families coding for channel proteins such as aquaporins might be potential candidates (Maurel et al., 2008; Wang et al., 2008,).

OsDUR3 transports urea with a high substrate affinity

This study is the first to identify and characterize a crop high-affinity urea permease. We functionally characterized OsDUR3 as urea transporter by three different approaches. First, expression of OsDUR3 complemented growth of the urea uptake-defective yeast mutant YNVW1 on 2 mM urea (Fig. 2), indicating a function of OsDUR3 in urea movement across the yeast PM. Second, OsDUR3 expression in Xenopus oocytes caused a significant accumulation of externally applied [14C]urea in the cells in a concentration-dependent and substrate-specific manner (Fig. 3a,c,d), again strongly supporting OsDUR3-mediated urea import. Finally, OsDUR3 expression restored the growth of the atdur3-1 mutant (Fig. 7a, b), where the active urea/proton symporter is disrupted (Kojima et al., 2007). On low urea, increased uptake of urea from medium was clearly observed in OsDUR3-(over)expressing Arabidopsis lines (Fig. 7c), pointing to a physiological role of OsDUR3 in urea transport associated with N nutrition or urea-N utilization in planta.

The transport kinetics of OsDUR3 for urea was successfully determined with OsDUR3 cRNA-injected oocytes. The import was saturable at c. 80 μM urea and conformed to a Michaelis–Menten kinetics (Fig. 3a) with an apparent substrate-affinity constant (Km) at c. 10 μM, which is in a similar lower-micromolar range, as determined for its orthologs from Arabidopsis thaliana (AtDUR3, Km of 3 μM, Liu et al., 2003a,b) and Paxillus involutus (PiDUR3, Km of 31.8 μM; Morel et al., 2008). Because the Km value of OsDUR3 for urea import into oocytes is consistent with the Km of c. 8 μM for urea absorption by rice roots (Fig. 1b), and because no additional high-affinity urea permease could be conspicuously predicted from the fully-sequenced rice genome, OsDUR3 most likely represents a main genetic component responsible for high-affinity urea transport system in rice. Thus, knowledge of OsDUR3 contributes to our understanding of rice membrane proteins required for N transport and nutrition.

Energetically, as the presence of more protons (e.g. at medium pH5) clearly favored urea influx into OsDUR3-expressing oocytes (Fig. 3b) and as growth complementation of a yeast mutant by OsDUR3 occurred better at low pH (Fig. 2), we suggest that OsDUR3 is a secondary active transporter that likely uses an electrochemical potential as a driving-force for urea movement across the PM, similar to its ortholog AtDUR3 (Liu et al., 2003a). Also urea influx into rice roots was sensitive to the protonophore CCCP (Fig. 1a), a membrane-associated ATPase uncoupler, affirming the existence of at least one energy-dependent component contributing to the movement of urea at micromolar concentrations across the PM of intact rice root-cells.

The physiological role of OsDUR3 is associated with plant nitrogen nutrition

Transcriptional upregulation of membrane permeases for the transport of nutrients/solutes by roots appears to be a common feature of plant responses to nutrient limitation (Takahashi et al., 2000; Liu et al., 2003a; Nazoa et al., 2003). In fact, levels of OsDUR3 mRNA in rice roots increased significantly after 2 d of N starvation (Fig. 5a). The upregulation of OsDUR3 was further increased upon root urea-resupply after the N deficiency but not changed obviously upon NH4+-resupply (Fig. 5a). Thus, OsDUR3 expression was also substrate inducible. Such a N-dependent elevation of the transcript-abundance in the roots strongly indicates that OsDUR3 may represent an important factor required for rice responsive to N limitation and for root-specific acquisition of urea when occurring in environments, similar to AtNRT2;1 for nitrate and AtDUR3 for urea in Arabidopsis (Nazoa et al., 2003; Kojima et al., 2007). Similar patterns of OsDUR3 expression in the shoots (Fig. 5a) would indicate that OsDUR3 might be needed for urea mobilization within the plant.

Moreover, a possible physiological relevance of OsDUR3 in urea transport could be linked to plant N remobilization. In germinating seeds or senescing plant parts, N-containing molecules such as proteins and nucleotides are broken down (to produce, e.g. urea and ammonium) and subsequently remobilized to growing organs (Polacco & Holland, 1993). Concentrations of OsDUR3 mRNA markedly increase during seed germination at a later stage in particular (Fig. 6a), comparable to the expression fashion of OsAMT1;2 (identical to OsAMT1;3; Sonoda et al., 2003; Fig. 6a) and AtDUR3 (Liu et al., 2003a). Thus, this result implies involvement of OsDUR3 in the seedling response to N deficiency and a requirement of the transporter for urea-N mobilization for developing new organs during seedling growth. At the tissue level, expression of OsDUR3 was higher in old leaves than in young leaves and developing flowers (Fig. 6b), consistent with the finding of a higher urea content in the old leaves (Fig. 9d), again suggesting that OsDUR3 would play a role in the redistribution and reutilization of urea-N from senescing parts to growing tissues. If so, OsDUR3 might also be critical for vascular (e.g. phloem) urea loading and/or unloading. To test this hypothesis, future work may focus on certain molecular and physiological analyses, for example, survey of promoter activity, protein localization in intact rice and concentrations of N-containing compounds including urea in different tissues.

Construction and transient-expression of OsDUR3:GFP in Arabidopsis protoplasts showed that OsDUR3 was mostly targeted to the PM (Fig. 4). Thus, based on current knowledge, OsDUR3 could mediate urea transport from, for example, soil environments into the cells. Transgenic approaches were used to provide convincing evidence that OsDUR3 does fulfill a function in effective urea acquisition and use in planta. Introduction of OsDUR3-ORF into the mutant atdur3-1 or its corresponding wild-type enabled all OsDUR3-transformed lines to grow better on low urea as a sole N source compared with nontransformed ones (Fig. 7a,b), demonstrating a contribution of OsDUR3 in urea uptake as well as possibly improved urea-N utilization by OsDUR3 (over)expression in plants. This interpretation is further supported by a 20-min root [15N]urea influx study showing that OsDUR3 expression in Arabidopsis enhanced urea absorption by sixfold (Fig. 7c). More direct physiological evidence for OsDUR3 contributing to the uptake and utilization of urea in low-millimolar ranges comes from the analysis of a T-DNA insertion rice mutant osdur3-dT. When grown with urea, osdur3-dT plants showed reduced growth and N-deficiency symptoms (Fig. 9a,b). A marked decrease in urea uptake was measured in the mutant (Fig. 9c). Moreover, a greater accumulation of urea in the old leaf of osdur3-dT mutant than that of the wild-type implies an additional critical role for OsDUR3 in the remobilization and/or reuse of nitrogen in the form of urea during plant senescence. However, more intense assessment of OsDUR3 role in plant urea-N physiology may still need to be performed in future with transgenic rice lines, that is, OsDUR3-(over) expressed osdur3-dT and WT plants.

The present work provides molecular and physiological evidence that rice OsDUR3 functions as a vital genetic component responsible for high-affinity transport of urea, which normally occurs in plant cells and also serves as a N source available in (soil) environments for plant growth. OsDUR3 should offer a potential and promising target for biotechnological approaches aiming to improve N-fertilizer use efficiency in crops.


We thank Dr von Wirén (IPK, Gatersleben, Germany) for kindly providing the atdur3-1 mutant, Dr Schumacher (UH, Germany), Dr Ludewig (TUD, Germany) and Dr Boles (GUF, Germany) for gifts of plasmids (pHXT426, pCF203, pOO2). We also thank Dr Roy, Dr Cotsaftis and Dr Kaiser (ACPFG, Australia) for their kind help during W-HW’s 1-yr joint-work in Australia. This work was financially supported by the NHTRD-Program 863 of China (no. 2006AA10Z166), NSF of China (NSFC) (no. 30771288, 31070223), and The Innovative Group Grant (no. 30821003) of the NSFC.