Competition between uptake of ammonium and potassium in barley and Arabidopsis roots: molecular mechanisms and physiological consequences

Abstract

Plants can use ammonium (NH₄⁺) as the sole nitrogen source, but at high NH₄⁺ concentrations in the root medium, particularly in combination with a low availability of K⁺, plants suffer from NH₄⁺ toxicity. To understand the role of K⁺ transporters and non-selective cation channels in K⁺/NH₄⁺ interactions better, growth, NH₄⁺ and K⁺ accumulation and the specific fluxes of NH₄⁺, K⁺, and H⁺ were examined in roots of barley (Hordeum vulgare L.) and Arabidopsis seedlings. Net fluxes of K⁺ and NH₄⁺ were negatively correlated, as were their tissue concentrations, suggesting that there is direct competition during uptake. Pharmacological treatments with the K⁺ transport inhibitors tetraethyl ammonium (TEA⁺) and gadolinium (Gd³⁺) reduced NH₄⁺ influx, and the addition of TEA⁺ alleviated the NH₄⁺-induced depression of root growth in germinating Arabidopsis plants. Screening of a barley root cDNA library in a yeast mutant lacking all NH₄⁺ and K⁺ uptake proteins through the deletion of MEP1–3 and TRK1 and TRK2 resulted in the cloning of the barley K⁺ transporter HvHKT2;1. Further analysis in yeast suggested that HvHKT2;1, AtAKT1, and AtHAK5 transported NH₄⁺, and that K⁺ supplied at increasing concentrations competed with this NH₄⁺ transport. On the other hand, uptake of K⁺ by AtHAK5, and to a lesser extent via HvHKT2;1 and AtAKT1, was inhibited by increasing concentrations of NH₄⁺. Together, the results of this study show that plant K⁺ transporters and channels are able to transport NH₄⁺. Unregulated NH₄⁺ uptake via these transporters may contribute to NH₄⁺ toxicity at low K⁺ levels, and may explain the alleviation of NH₄⁺ toxicity by K⁺.

Introduction

Ammonium (NH₄⁺) is a central nitrogen compound in all organisms. In the autotrophic anabolism of plants, NH₄⁺ is generated through the reduction of nitrate, or is directly taken up from the soil. In addition, NH₄⁺ is generated via de-amination of organic compounds during protein turnover, photorespiration, and lignin biosynthesis (Joy, 1988).

It is well known that high concentrations of NH₄⁺ can be toxic to plants leading to severe growth depression (Gerendás et al., 1997). Various hypotheses have been put forward aimed at identifying the cause of NH₄⁺ toxicity. The majority of these hypotheses deal with the physiological changes associated with NH₄⁺ assimilation and ion imbalances resulting from decreased uptake of essential cations such as K⁺, Mg²⁺, and Ca²⁺ (Barker et al., 1967; Gerendás et al., 1997; Roosta and Schjoerring, 2007). Rapid assimilation of NH₄⁺ in roots is associated with a large requirement for carbohydrates and may, consequently, result in root carbon depletion (Kronzucker et al., 1998; Schjoerring et al., 2002). NH₄⁺ assimilation also leads to a net release of H⁺ and a subsequent reduction in the concentration of dicarboxylic acids by decarboxylation, a cellular response to maintain cytoplasmic pH (Raven and Smith, 1976). Decarboxylation of dicarboxylic acids puts a further strain on the carbon budget of the root, and results in an increased uptake of anions for charge balance. In addition, NH₄⁺ toxicity has been associated with hormonal imbalances (Barker, 1999; Walch-Liu et al., 2000) and a decrease in photosynthesis (Putrich and Barker, 1967; Gerendás et al., 1997).

Recently, an alternative hypothesis was put forward suggesting that growth depression of barley plants at high NH₄⁺ was due to a futile cycling of NH₄⁺ across the plasma membrane, where active extrusion of excess NH₄⁺ by an as yet unknown transporter led to extensive energy consumption (Britto et al., 2001). The high respiration rate of roots in conjunction with the energy demand for active extrusion of NH₄⁺ was suggested to be the fundamental cause of the damaging effects of high external NH₄⁺. Thus, toxicity was explained by an excessive cycling at the plasma membrane as opposed to consequences of assimilation and accumulation in the cells. Several bacteria also show elevated respiration at high NH₄⁺ and low K⁺ levels, caused by the energetically unfavourable cycling of NH₄⁺ at the plasma membrane (Buurman et al., 1989, 1991). It was suggested that cycling of NH₄⁺ across bacterial plasma membranes is caused by the disruption of the regulation of NH₄⁺ uptake and non-specific influx through the kdp K⁺ transporter (Buurman et al., 1991).

It has long been known that toxicity of NH₄⁺ is particularly pronounced during K⁺ deficiency (Wall, 1939). Similarly, addition of K⁺ can alleviate NH₄⁺ toxicity (Barker et al., 1967; Cao et al., 1993). Toxicity of NH₄⁺ and alleviation by K⁺ has also been shown in yeast, where high concentrations of NH₄⁺ are toxic only when K⁺ levels are low (Hess et al., 2006).

NH₄⁺ and K⁺ are highly similar regarding charge, size, and hydration energy, characteristics that are important for membrane transport (Wang et al., 1996; White, 1996). NH₄⁺ has been shown to influence the uptake and accumulation of K⁺ (see, for example, Rufty et al., 1982; Allen and Raven, 1987; van Beusichem et al., 1988; Finnemann and Schjoerring, 1999; Angeles Martinéz-Cordero et al., 2005; Szczerba et al., 2008b, and references therein), and K⁺ influences the uptake and accumulation of NH₄⁺ (Scherer et al., 1984; Nielsen and Schjoerring, 1998; Szczerba et al., 2008a). Reduced NH₄⁺ uptake and accumulation through the provision of K⁺ may prevent competition with NH₄⁺ in subsequent metabolic processes that require K⁺. It is therefore likely that the toxicity of NH₄⁺ at high external concentrations is closely linked to its interference with K⁺ uptake and cellular homeostasis.

The genome of Arabidopsis thaliana contains around 20 genes encoding K⁺-selective transporters and 57 genes encoding cation channels (Mäser et al., 2001; Véry and Sentenac, 2002). The presence of the non-selective cation channel (NSCC) blocker lanthanum (La³⁺) reduced the uptake of NH₄⁺ in leaves and roots (Nielsen and Schjoerring, 1998; Szczerba et al., 2008a) and several experiments have indicated that NSCCs can facilitate the transfer of NH₄⁺ across membranes (Gassman and Schroeder, 1994; White, 1996; Davenport and Tester, 2000; Demidchik and Tester, 2002; Balleza et al., 2005). In addition, experiments using different heterologous expression systems have suggested that the Arabidopsis K⁺ channel AtKAT1 (Schachtman et al., 1992; Bertl et al., 1995; Cao et al., 1995; Uozumi et al., 1995; Moroni et al., 1998), K⁺ transporters AtKCO1 (Czempinski et al., 1997) and HvHKT2;1 (Santa-María et al., 2000) are also candidates for mediating NH₄⁺ transport.

The aim of this work was to investigate the effect of different concentrations of NH₄⁺ and K⁺ on the growth of barley (Hordeum vulgare L.) and Arabidopsis seedlings, and to demonstrate the interdependence of K⁺ and NH₄⁺ fluxes during uptake. Pharmacological treatments with TEA⁺ and Gd³⁺ supported the indication from earlier studies that K⁺-selective channels as well as NSCCs transport NH₄⁺ in the roots. Strikingly, treatment with TEA⁺ alleviated NH₄⁺-induced growth depression in germinating Arabidopsis seedlings, indicating a role for K⁺-selective channels in mediating NH₄⁺ uptake. A yeast complementation screen using a cDNA library from barley roots indicated that HvHKT2;1 transports NH₄⁺, and that growth on NH₄⁺ was inhibited by increasing K⁺ concentrations. Similar results were seen for AtAKT1 and AtHAK5. K⁺ transport through HvHKT2;1, AtAKT1, and AtHAK5 was inhibited by NH₄⁺. Together these results demonstrate interference between K⁺ and NH₄⁺ at uptake via HvHKT2;1, AtAKT1 and AtHAK5.

Materials and methods

Plant cultivation

Arabidopsis seeds (Col-0, Lehle seeds) were surface-sterilized by washing for 1 min with 50% ethanol, followed by incubation in 5% NaOCl and 0.02% SDS for 10 min. Thereafter, seeds were rinsed five times with sterile, double-deionized water. They were submerged in ±0.05% agarose and stratified for 2 d at 4 °C, either in the agarose-solution or on media in square 50 ml Petri dishes with the same composition as described by Cao et al. (1993). The medium was supplied with 0.4% agarose and 1% sucrose, and different combinations of 0.2 mM KCl, 6 mM NH₄Cl, 1 mM GdCl₃ or 10 mM TEACl. On the top of each plate two rows of seeds were sown, with 30 seeds in each row. Plates were organized horizontally, allowing seedlings to grow along the surface of the medium, in a controlled environment growth-chamber with a 85–110 μmol m⁻² s⁻¹ photon flux density for 8 h d⁻¹, 75–80% air humidity, and 20 °C air temperature. Plants were cultured for 10 d prior to analysis.

Barley seeds (Hordeum vulgare L., cv. Antonia) were sown on vermiculite in a greenhouse (250–300 μmol m⁻² s⁻¹ photon flux density, 75–80% humidity, at 20 °C and 18 °C during 16/8 h light/dark, respectively). Seedlings were watered with (K⁺-free) double-deionized water. When the shoot was about 8 cm, four seedlings were transferred to 4.0 l buckets containing hydroponic medium without K⁺ or NH₄⁺. The medium consisted of 0.3 mM MgSO₄-7H₂O, 0.1 mM NaCl, 0.2 mM NaH₂PO₄-H₂O, 0.2 mM Na₂SO₄-10H₂O, 0.15 mM Mg(NO₃)₂-6H₂O, 0.6 mM Ca(NO₃)₂-4H₂O, 1.5 mM Fe(III)-EDTA-Na, 150 μM MnSO₄-H₂O, 105 μM ZnSO₄-7H₂O, 120 μM CuSO₄-5H₂O, 300 μM H₃BO₃, and 120 μM Na₂MoO₄-2H₂O. The pH was kept at 6.0 with 1.2 mM MES/NaOH. The solution was aerated with filtered air supplied through a syringe needle. After 3 d of growth in hydroponic medium, various concentrations of NH₄⁺ and K⁺ were added using (NH₄)₂SO₄ and/or K₂SO₄. The nutrient solution was changed daily. Five days after the initiation of the K⁺/NH₄⁺ treatments, plants were harvested in a mixed sequence to reduce artefacts caused by circadian rhythms. For NH₄⁺ and K⁺ determination, roots were washed in 2.5 g l⁻¹ CaSO₄, and three times in double-deionized water, then dried on paper tissue. Plants were fractionated into (i) oldest leaf, (ii) young leaves, (iii) stems, and (iv) roots. Each fraction was weighed, frozen in liquid nitrogen, and stored at –80 °C prior to extraction.

Root length measurements

Arabidopsis roots were magnified and measured using a scanner or a microscope (Leica Microsystems M500 or MZ FL III) connected to a camera (Leica Microsystems DC 300 F).

Measurement of soluble NH₄⁺ content

Samples were homogenized with a steel pin, after which 200 mg was taken out and crushed with 100 μl acid-washed sand and 2 ml 10 mM ice-cold formic acid, in an agate mortar on ice. This suspension was centrifuged at 10 000 g and 4 °C. The supernatant was centrifuged again and the supernatant of this last centrifugation-step was diluted ten times with 10 mM formic acid. The NH₄⁺ concentration was determined by measuring the fluorescence in 3 mM o-phthalaldehyde, 10 mM 2-mercaptoethanol, and 100 mM phosphate buffer adjusted to pH 6.8 as previously described (Husted et al., 2000).

Potassium determination with ICP-OES

Samples from the NH₄⁺-extraction (described in the previous paragraph) were diluted ten times with 7% HNO₃. The K⁺ concentration was determined by Induced Coupled Plasma-Optical Emission Spectroscopy (ICP-OES; Perkin-Elmer Optima 3000XL) at axial mode using a wavelength of 766.474 nm.

Flux measurements

Net fluxes of NH₄⁺, K⁺, and H⁺ at the surface of Arabidopsis roots were measured non-invasively using the MIFE^® technique (UTas Innovation, Hobart, Australia) essentially as described by Shabala et al. (1997, 2001) and Shabala (2000). Briefly, microelectrodes were pulled and salinized with tributylchlorosilane. After backfilling with an appropriate electrolyte, electrode tips were filled with commercially available ionophore cocktails (09882 for NH₄⁺; 60031 for K⁺; 95297 for H⁺; all from Fluka) and mounted on a 3D-micromanipulator (MMT-5; Narishige), with their tips close together, 20 μm above the root surface. During measurements, a computer-controlled stepper motor moved the electrodes between two positions (20 μm and 40 μm, respectively) from the root surface in a 10 s square-wave manner. The CHART software (Shabala et al., 1997; Newman, 2001) recorded the potential difference between two positions and converted them into electrochemical potential differences using the calibrated Nernst slope of the electrodes. H⁺ fluxes were then calculated using the MIFEFLUX software for cylindrical diffusion geometry (Newman, 2001). NH₄⁺ and K⁺ fluxes were calculated using the algorithm described in Knowles and Shabala (2004) taking into account the mutual interference between the two ions resulting from non-ideal selectivity of both the K⁺ and NH₄⁺ ionophore cocktails.

Arabidopsis Col-0 plants were pre-grown on plates for 10 d as described. Eight to 10-mm-long apical root segments were excised and mounted horizontally in a Perspex holder using agar (see Babourina et al., 2000, for details). The holder was immediately placed in a 4 ml measuring chamber filled with 0.1 mM CaCl₂, pH 5.7 (unbuffered), mounted on a computer-driven 3D-micromanipulator (MMT-5; Narishige) and left to equilibrate for 60 min before the addition of treatment. Steady-state ion fluxes were recorded over a period of 5 min, 15 min after the application of the appropriate treatment. Unless stated differently, 1 mM NH₄⁺, 0.2 mM K⁺, 10 mM TEA, and 0.1 mM Gd³⁺ were used in the various treatments.

Yeast strains and expression of plant transporters

Three deletion mutants in the yeast Saccharomyces cerevisiae were generated in this study: a triple NH₄⁺ transporter deletion mutant Δmep1-3, a K⁺ transporter double deletion mutant Δtrk1,2 and a 5-fold deletion mutant Δmep1-3 Δtrk1,2. All mutants were made in the BY background from the Euroscarf collection. To generate Δmep1-3, HIS3 was first generated by PCR using primers: fw 5′-CACATGTGCTAACCAAACATCAGTGGGTAGTAATCATTCGGCGTCAGCGGGTGTTGG-3′ and rev 5′-CGTTGGCATGCGATGAGGTCAGTTTCCTGCGCCATTTCAGTCTCGGTCTATTCTTTTGATTTAT-3′ to form a deletion cassette for MEP2. The PCR product was transformed into Δmep1 (Y04751, Mata his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 YGR121c::kanMX4), thus yielding a double mutant Δmep1,2. The double mutant was crossed with Δmep3 (Y15553, Matα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 YPR138c::kanMX4) to yield Δmep1-3 (Matα leu2Δ0 lys2Δ0 ura3Δ0 YGR121c::kanMX4 YNL142w::HIS3 YPR138c::kanMX4). Yeast strain Δtrk1,2 (Mata his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 YKR050w::kanMX4, YJL129c::kanMX4) was generated by crossing Δtrk1 (Y15121 Matα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 YKR050w::kanMX4) and Δtrk2 (Y01296 Mata his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 YJL129c::kanMX4). Likewise, the 5-fold deletion mutant strain Δmep1-3 trk1,2 (Mata leu2Δ0 ura3Δ0 YGR121c::kanMX4 YNL142w::HIS3 YPR138c::kanMX4, YKR050w::kanMX4, YJL129c::kanMX4) was generated by crossing the two aforementioned yeast strains Δmep1-3 and Δtrk1,2. All crossings (mating and sporulation) were done as described by Kassir and Simchen (1991). The deletion of genes was successfully verified by PCR, using a reverse primer for the flanking kanamycin gene (rev 5′-CTAGTAATCTTCAGGGGCC-3′) in combination with forward primers for the genes: MEP1 5′-ATGGCAACATCAGCTCG-3′, MEP3 5′-ATGGCTACATCTGCTAGAAGA-3′, TRK1 5′-TTTCACGAAAAGAGGGACAATGTAC-3′, and TRK2 5′-AGGATTCGTTGTGCTTGTTGAATCG-3′. The genotypes of the resulting yeast strains were characterized by cultivation on different amino acids and NH₄⁺ and K⁺ conditions and their mating type was determined by mating with haploid strains with supplementing marker-genes and known mating type. To enable selection using two markers only, and to grow cells on media without supplementation by amino acids, yeast strains were transformed with single copy vectors from the pRS series containing appropriate marker genes.

Yeast strains were transformed with pYES2 containing cDNAs encoding Arabidopsis hak5L₇₇₆H (Rubio et al., 2000), AtAKT1 in plasmid pFL61 (Sentenac et al., 1992) or a barley root cDNA library (Pedas et al., 2008). Transformation was done by electroporation, as described at http://www.agr.kuleuven.ac.be/dp/logt/protocol/yeastelectroporation.htm. After transformation, yeast cells were pre-grown on synthetic YNB medium without NH₄⁺, supplemented with 100 mM K₂SO₄ and 0.1% proline, on 2% glucose and a pH of 5.5 (50 mM succinic acid/TRIS). Growth assays were carried out on synthetic medium containing 1% galactose, YNB without amino acids, K⁺ and NH₄⁺ and supplemented with various concentrations of (NH₄)₂SO₄ or K₂SO₄ as indicated. Plates were incubated at 30 °C for at least 3 d. The K⁺ content of agar was determined to be 500 μg g⁻¹, which has been accounted for in the composition of the media.

Statistical analysis

Statistical analysis was performed using SAS software (SAS Institute Inc.,USA, version 9.1).

Results

Uptake and accumulation of NH₄⁺ and K⁺ in barley

To characterize potential interferences in the uptake and accumulation of NH₄⁺ and K⁺ in barley, seedlings were cultivated in hydroponic medium, containing NO₃⁻ but neither NH₄⁺ nor K⁺. Subsequently, seedlings were transferred for 5 d to medium supplemented with various NH₄⁺ and K⁺ concentrations. Fresh weight and the soluble NH₄⁺ and K⁺ content were determined.

Cultivation for 5 d on high levels (3–6 mM) of NH₄⁺ without K⁺, significantly decreased both root and shoot fresh weight compared with plants grown in similar media but in the absence of NH₄⁺ (Figs 1, 2). This growth depression by NH₄⁺ was progressively alleviated by the presence of sub-millimolar concentrations of K⁺. The weight of seedlings grown in the absence of K⁺ and NH₄⁺ was only slightly decreased compared with plants that had received K⁺, but no NH₄⁺ (Fig. 2). Three days on 6 mM NH₄⁺ and 0.2 mM K⁺, followed by a 2 d K⁺-starvation period with unchanged NH₄⁺ provision did not produce any significant difference in fresh weight relative to plants with constant K⁺ supply.

Fig. 1.

Barley plants (cv. Antonia) cultivated at various K⁺ and NH₄⁺ concentrations. Plants were grown in a greenhouse in hydroponic medium with all nutrients including 1.5 mM NO₃⁻ as nitrogen source but without NH₄⁺ or K⁺ for 3 d and subsequently transferred for 5 d to various K⁺ and NH₄⁺ concentrations as indicated.

Fig. 2.

Fresh weight of barley plants (cv. Antonia) cultivated at various K⁺ and NH₄⁺ concentrations. Plants were grown in a greenhouse in hydroponic medium with all nutrients including 1.5 mM NO₃⁻ as nitrogen source but without NH₄⁺ or K⁺ for 3 d and subsequently transferred for 5 d to various K⁺ and NH₄⁺ concentrations as indicated. Black bars represent root weight; white bars represent shoot weight, with old leaves (bottom), young leaves (middle), and stem (top). Data represent means ±SE of 12 plants for each measurement.

The soluble NH₄⁺ content (Fig. 3A) of plants cultivated in the absence of NH₄⁺ was in the range of 0.4–0.7 μmol g⁻¹ FW in all tissues, regardless of the K⁺ concentration in the medium. A concentration of 0.4–0.7 μmol NH₄⁺ g⁻¹ FW may therefore reflect the basal NH₄⁺ concentration resulting from nitrate reduction and processes such as protein turnover and photorespiration. The presence of NH₄⁺ in the nutrient solution led to elevated levels of soluble NH₄⁺ in all tissues, with maxima of 5.6 μmol g⁻¹ FW in young leaves and 22.9 μmol g⁻¹ FW in roots when NH₄⁺ was supplied at 6 mM. These amounts were reduced in the presence of K⁺, with a higher effect in shoot than in root tissue (Fig. 3A).

Fig. 3.

NH₄⁺ (A) and K⁺ (B) tissue concentration of barley plants (cv. Antonia) cultivated on different NH₄⁺ and K⁺ regimes in hydroponic medium. Plants were grown in hydroponic medium including all nutrients and 1.5 mM NO₃⁻ as nitrogen source but without NH₄⁺ or K⁺ for 3 d and subsequently transferred to similar medium including various K⁺ and NH₄⁺ concentrations (as indicated) for an additional 5 d. Data represent means ±SE of three measurements with material from four plants per measurement.

The corresponding K⁺ content showed values as low as 10 μmol g⁻¹ FW in roots and 20 μmol g⁻¹ FW in leaves of plants that had never received any K⁺ other than their seed reserves (Fig. 3B). In these conditions, increasing levels of NH₄⁺ in the growth media had a negative effect on the K⁺ concentrations of all the tissues tested, in line with the interpretation that NH₄⁺ causes K⁺ extrusion (Szczerba et al., 2006). This reduction in K⁺ concentration was particularly apparent in roots, where a concentration as low as 3.9 μmol K⁺ g⁻¹ FW was observed in plants that were not supplied with K⁺. When neither K⁺ nor NH₄⁺ was supplied, the K⁺ concentration of roots was more than twice as high (11.5 μmol g⁻¹ FW). In the absence of external NH₄⁺, the addition of 0.2, 0.4 or 1.2 mM K⁺ resulted in comparable levels of K⁺, approximately 50 μmol g⁻¹ in roots and 140 μmol g⁻¹ in leaves. Withholding K⁺ during the final 2 d of a treatment with 6 mM NH₄⁺ and 0.2 mM K⁺ reduced the K⁺ concentration in roots, but not in leaves, indicating that barley plants are very efficient in controlling and retaining K⁺ in leaves.

Besides the smaller size and chlorotic appearance of the young leaves of plants grown on high NH₄⁺ concentrations, in the absence of K⁺, there were no clear phenotypic differences between plants in different treatments. In the presence of low K⁺ concentrations, the negative effects of NH₄⁺ provision were ameliorated (Fig. 1).

Blocking K⁺ uptake of Arabidopsis seedlings by pharmacological inhibitors

An earlier study by Cao et al. (1993) reported the dramatic depression of root growth of Arabidopsis seedlings caused by the presence of NH₄⁺ in the absence of K⁺ in the growth media. To study the effect reported by Cao et al. (1993) further, Arabidopsis seeds were germinated on media containing 6 mM NH₄⁺ in the absence of K⁺. Seedlings hardly developed roots (Fig. 4). The addition of 0.2 mM K⁺ to the 6 mM NH₄⁺ medium restored root growth above the level measured when neither NH₄⁺ nor K⁺ were present, essentially confirming the earlier results of Cao et al. (1993). To investigate whether the amelioration of NH₄⁺ toxicity by K⁺ was due to competition between K⁺ and NH₄⁺ at the same uptake site, 10 mM tetraethyl ammonium (TEA⁺), a known blocker of K⁺ channels (Yellen et al., 1991; Lenaeus et al., 2005), was added to the medium (Fig. 4). Addition of TEA⁺ generally led to a substantial, significant reduction in root growth, which was most pronounced in the absence of K⁺ in the growth medium. This pronounced root growth depression may be due to the inhibition of remobilization of K⁺, which had initially derived from seed reserves. However, growth depression by 10 mM TEA⁺ in the absence of K⁺ and NH₄⁺ was less pronounced than growth depression by the application of 6 mM NH₄⁺ alone, the latter of which almost completely inhibited root formation. More importantly, in the presence of 10 mM TEA⁺ plus 6 mM NH₄⁺, growth was not significantly different from growth in the presence of 10 mM TEA⁺ alone (Figs 4, 5). In fact, in this case the addition of TEA⁺ significantly improved growth and completely abolished the NH₄⁺-induced growth depression. These data suggest that K⁺ efficiently reduced the uptake of NH₄⁺, but that NH₄⁺ toxicity was not exclusively the result of competition with K⁺. Rather, NH₄⁺ toxicity seemed to directly relate to an increased flux of NH₄⁺ into the roots. Addition of Gd³⁺, a known blocker of NSCCs (Demidchik et al., 2002), had a minor effect and did not reverse the effects of NH₄⁺ toxicity to the same extent as TEA⁺ (data not shown).

Fig. 4.

Primary root length of Arabidopsis Col-0 seedlings cultivated for 10 d on plates containing different amounts of NH₄⁺, K⁺, and TEA⁺ as indicated. Inset: Root length at treatments with high levels of NH₄⁺ and/or TEA⁺ in the absence of K⁺. Data represent means ±SE. Between 21 and 35 roots were measured per treatment.

Fig. 5.

Roots of Arabidopsis Col-0 plants grown on medium without K⁺, but in the presence of 6 mM NH₄⁺ (C, D) or 6 mM NH₄⁺ plus 10 mM TEA⁺ (A, B). (This figure is available in colour at JXB online.)

Fluxes of NH₄⁺ and K⁺ under the influence of uptake blockers in Arabidopsis

To characterize more closely the interference between NH₄⁺ and K⁺ at the site of uptake at the root surface, net NH₄⁺, K⁺, and H⁺ fluxes were determined in the mature root zone of Arabidopsis seedlings with microelectrodes using the MIFE^® technique (Shabala et al., 1997; Newman, 2001). Roots of seedlings grown on agarose plates without K⁺ and NH₄⁺ (as described previously), were exposed to various bathing solutions. When fluxes of NH₄⁺ and K⁺ were measured at constant 0.2 mM K⁺ and various NH₄⁺ concentrations (Fig. 6), an almost linear inverse correlation of fluxes for the two cations was found. Increasing NH₄⁺ in the solution increased the NH₄⁺ influx, while reducing the K⁺ influx. Addition of NH₄⁺ and K⁺ separately resulted in increased NH₄⁺ and K⁺ influxes, respectively (Fig. 7). Interestingly, in the presence of both K⁺ and NH₄⁺, net influx of NH₄⁺ was about double that measured upon sole NH₄⁺ exposure. Conversely, net influx of K⁺ was reduced to about half of the flux measured under sole K⁺ exposure. Thus, K⁺ stimulated NH₄⁺ influx, while NH₄⁺ reduced the influx of K⁺. The high net influx of NH₄⁺ in the presence of both ions was accompanied by a marked increase in net H⁺ efflux across the plasma membrane. The addition of 10 mM TEA⁺ and 0.1 mM Gd³⁺ strongly reduced both NH₄⁺ and K⁺ influxes, supporting the view that stimulation of NH₄⁺ influx by K⁺ and the inhibition of K⁺ influx by NH₄⁺ were related to one or several common uptake systems. Application of TEA⁺ did not affect H⁺ extrusion in the absence of K⁺ and NH₄⁺, but reduced H⁺ effluxes in solution where K⁺ and NH₄⁺ were present. Net H⁺ efflux was less affected by Gd³⁺ than by TEA⁺ (Fig. 7).

Fig. 6.

Net fluxes of NH₄⁺ and K⁺ at the root surface are dependent on the external concentration of NH₄⁺. Arabidopsis Col-0 plants were grown on agarose plates with NO₃⁻ as nitrogen source and without K⁺ and NH₄⁺. Measurements were done at constant external K⁺ of 0.2 mM in bathing solution (0.1 mM CaCl₂).

Fig. 7.

Net fluxes of NH₄⁺ (A), K⁺ (B), and H⁺ (C). Arabidopsis Col-0 plants were grown on agarose plates with NO₃⁻ as nitrogen source and without K⁺ and NH₄⁺. Fluxes were measured for 5 min in bathing solution (0.1 mM CaCl₂) 15 min after the addition of various NH₄⁺, K⁺, TEA⁺, and Gd³⁺ as indicated.

NH₄⁺/K⁺ interference via plant transporters expressed in yeast

The results described above suggest that certain plasma membrane transporters or channels in both Arabidopsis and barley are capable of transporting NH₄⁺ and K⁺. To identify putative transporters or channels at the molecular level, several new mutants lacking multiple NH₄⁺ and K⁺ transporters were generated from single deletion mutants in the yeast BY4741 background (Euroscarf): (i) an NH₄⁺ transporter mutant Δmep1-3, devoid of all three NH₄⁺ transporter genes MEP1, MEP2, and MEP3, (ii) a K⁺ transporter mutant Δtrk1,2 lacking the two K⁺ transporter genes TRK1 and TRK2, and (iii) a 5-fold deletion mutant Δmep1-3 Δtrk1,2 lacking all five genes. The triple mutant Δmep1-3, not expressing the NH₄⁺ transporters Mep1p, Mep2p, and Mep3p, did not grow on media containing 10 mM NH₄⁺ as the sole nitrogen source after 3 d of incubation at 30 °C (Fig. 8A, empty vector control). The trk1,2 double mutant did not grow at less than 10 mM K⁺ in the medium when proline was the sole nitrogen source (Fig. 8B, empty vector control). Deletion of all five genes in Δmep1-3 Δtrk1,2 thus provided a tool for the identification of transporters that transport both NH₄⁺ and K⁺ at limiting concentrations.

Fig. 8.

Growth assays of various yeast strains transformed with cDNA encoding HvHKT2;1, ΔN₆₂-hkt2;1, AtAKT1, or hak5L₇₇₆H at various K⁺ and NH₄⁺ concentrations or 0.1% proline instead of NH₄⁺ as nitrogen source. The various yeast mutant strains were (A) Δmep1-3, (B) Δtrk1-2, and (C) Δmep1-3 Δtrk1-2. Transformants were pre-grown on SD medium containing 0.1% proline as nitrogen source and supplemented with 100 mM K⁺. Yeast suspensions with an OD_600nm of 0.01 (upper drop) and 0.0001 (lower drop) were placed onto SG medium (1% galactose; YNB without amino acids, K⁺ and NH₄⁺) with various supplementations as indicated. Control, empty vector pYES2.

The yeast mutant Δmep1-3 Δtrk1,2 was transformed with a cDNA library made from roots of barley seedlings (Pedas et al., 2008). In the initial screen, transformants were grown on media containing six different combinations of pH, [NH₄⁺], and [K⁺]: pH 4.5 with 0.5 mM K⁺ and 5 or 10 mM NH₄⁺, pH 4.5 with 5 mM K⁺ and 5 or 10 mM NH₄⁺, pH 5.5 with 1 mM K⁺ and 20 mM NH₄⁺, and pH 5.5 with 5 mM K⁺ and 50 mM NH₄⁺. A total of 112 colonies from all conditions were harvested and plasmids of 24 of these transformants (four of each of the six conditions) were extracted, transformed into E. coli and subjected to DNA sequencing. A database search revealed that all 24 coding sequences were identical to HvHKT2;1. Strikingly, about half of the plasmids contained a version of HvHKT2;1 cDNA lacking either 36 or 85 nucleotides in the 5′ end when compared to full-length clones. Twenty nucleotides of this deletion represented a sequence of the 5′-UTR. An in-frame ATG codon, 186 nucleotides downstream of the original ATG start codon, probably served as a translational initiation point resulting in ΔN₆₂-hkt2,1 in both truncated versions. HvHKT2;1 and ΔN₆₂-hkt2;1 as well as hak5L₇₇₆H and AtAKT1 from Arabidopsis were subsequently characterized in the triple mutant Δmep1-3, the double mutant Δtrk1,2 and the 5-fold mutant Δmep1-3 Δtrk1,2 on media containing various NH₄⁺ and K⁺ concentrations (Fig. 8). The mutation L₇₇₆H in the AtHAK5 protein was earlier shown to be important for functional expression of AtHAK5 in yeast (Rubio et al., 2000).

Full-length HvHKT2;1 and AtAKT1 very weakly increased growth of Δmep1-3 compared with the empty vector control and other transporter genes at low NH₄⁺ (10 mM) and low K⁺ (0.15 and 1 mM; Fig. 8A). In all transformants of Δmep1-3, growth was suppressed by increasing K⁺ concentration irrespective of the NH₄⁺ level, indicating that K⁺ competed with NH₄⁺ (Fig. 8A) When transformed into Δtrk1,2, AtAKT1, HvHKT2;1, ΔN₆₂-hkt2;1, and hak5L₇₇₆H clearly complemented the growth deficiency of the mutant at low K⁺ concentrations (Fig. 8B). However, increasing NH₄⁺ at low K⁺ caused growth repression, particularly in Δtrk1,2 transformed with ΔN₆₂-hkt2;1, and hak5L₇₇₆H but also, albeit to a smaller extent, in Δtrk1,2 transformed with HvHKT2;1 and AtAKT1 (Fig, 8B) In the 5-fold mutant lacking all yeast endogenous NH₄⁺ and K⁺ transporters, only HvHKT2;1, AtAKT1 and, less markedly, hak5L₇₇₆H, complemented the growth deficiency of the mutant (Fig. 8C). Interestingly, when transformed with full-length HvHKT2;1, the triple mutant Δmep1-3 and the 5-fold mutant Δmep1-3 Δtrk1,2 grew well (Fig. 8C), whereas the double mutant Δtrk1,2 did not (Fig. 8B). The opposite was the case for its N-terminally truncated counterpart ΔN₆₂-hkt2;1, which complemented growth deficiency on low [K⁺] better in Δtrk1,2 than in Δmep1-3 or Δmep1-3 Δtrk1,2.

Discussion

Changes in cation composition, specifically a reduction in tissue levels of K⁺ have been reported in connection with NH₄⁺ toxicity in plants (see for example Scherer et al., 1984; Vale et al., 1987, 1988; Finnemann and Schjoerring, 1999; Santa María et al., 2000; Kronzucker et al., 2003; Szczerba et al., 2006). Elevated external concentrations of K⁺ can alleviate the effects of NH₄⁺ toxicity, and cause a decrease in the tissue concentrations of NH₄⁺ (Barker et al., 1967; Cao et al., 1993; Spalding et al., 1999; Szczerba et al., 2008a, and references therein). To understand the role of cation transporters and channels in NH₄⁺ toxicity better, plant growth, NH₄⁺ and K⁺ accumulation, and the specific fluxes of NH₄⁺, K⁺, and H⁺ in Arabidopsis and barley seedlings were examined. High NH₄⁺ and low K⁺ concentrations caused chlorosis in leaves and reduced biomass and root growth of barley plants (Figs 1, 2), suggesting that these conditions caused NH₄⁺ toxicity. All of these symptoms were alleviated by an elevated provision of K⁺.

In barley, plant-soluble NH₄⁺ content was low at high external K⁺ concentration and vice versa (Fig. 3; see also Szczerba et al., 2006, 2008a). In young leaves, however, K⁺ and NH₄⁺ concentrations were relatively constant, with NH₄⁺ only reaching high levels when the external NH₄⁺ concentration was very high and that of K⁺ very low. K⁺ levels in young leaves were always above 70–80 μmol g⁻¹ FW, unless K⁺ was absent from the growth medium. This constant level of K⁺ may reflect the important role of K⁺ in vital processes such as stomatal opening and closure (Humble and Hsiao, 1969; Cakmak et al., 1994; Dietrich et al., 2001) and protein and starch synthesis (Marschner, 1995). In K⁺-deficient barley plants several transport mechanisms may be activated to enable a more efficient translocation from root to shoot, thus sustaining the K⁺ concentration of the latter and securing K⁺-dependent metabolic processes (Britto and Kronzucker, 2006).

The exact mechanism behind the interaction of K⁺ and NH₄⁺ is not known, but it has been suggested that membrane passage of NH₄⁺ via K⁺ channels and NSCCs could play a role (Howitt and Udvardi, 2000; Kronzucker et al., 2001; Szczerba et al., 2008a, b). Our results show that the growth inhibition by high external NH₄⁺ was reduced in the presence of either TEA⁺ or Gd³⁺, which suggests a role for both K⁺-selective channels as well as NSCCs in mediating NH₄⁺ uptake. Interestingly, alleviation of NH₄⁺ induced growth depression by TEA⁺ was only seen when the growth medium was supplemented with sucrose (data from experiments without sucrose not shown). This may be explained by an inhibition of internal K⁺ transport by TEA⁺ causing a negative impact on carbohydrate and energy supply to the roots via the phloem which will be more severe in the absence of external sucrose.

To investigate more closely the transport of NH₄⁺ and K⁺ at the root surface, flux measurements were undertaken, using the MIFE^® technique (Newman, 2001). Net fluxes of NH₄⁺, H⁺, and K⁺ were recorded at the mature root zone of Arabidopsis seedlings. At a constant K⁺ concentration of 0.2 mM, net NH₄⁺ and K⁺ fluxes into the roots showed an almost linear inverse correlation with NH₄⁺ reducing the uptake of K⁺, and K⁺ reducing the uptake of NH₄⁺ (Fig. 6). This indicates competition for uptake sites between the two ions as also suggested by Scherer et al. (1984), Nielsen and Schjoerring (1998), and Szczerba et al. (2008a).

Although fluxes showed a reverse correlation, co-supplementing plants with K⁺ increased NH₄⁺ influx (Fig. 7). This could be the result of increased channel- or transporter activity to enable K⁺ uptake. Indeed, both Gd³⁺ and TEA⁺ strongly reduced the NH₄⁺ influx under these conditions with a combination of both blockers giving the largest effect. Very similar results were obtained in experiments using nutrient solution instead of CaCl₂ as the bathing medium (data not shown). The inhibition of NH₄⁺ influx by uptake blockers suggests that K⁺ channels/transporters as well as NSCCs are involved in membrane passage of NH₄⁺. This result is in contrast to reports by Szczerba et al. (2008a) who observed an increased NH₄⁺ influx at 0.1 and 40 mM K⁺ in the presence of TEA⁺, and concluded that the K⁺-sensitive influx of NH₄⁺ was mainly via the NSCC pathway.

Unlike the short-term K⁺ stimulation of NH₄⁺ influx in Arabidopsis, K⁺ provision to barley inhibited NH₄⁺ accumulation over time (Fig. 3). The stimulation of NH₄⁺ influx by K⁺ in our measurements with Arabidopsis may be related to the growth conditions: for flux measurement, plants were germinated in the absence of both K⁺ and NH₄⁺. This would lead to an induction of transporters, such as HAK5, which is repressed by high NH₄⁺ (Rubio et al., 2008) but induced by low K⁺ supply. It would have been relevant to measure fluxes from plants that were grown in the presence of high NH₄⁺ and in the absence of K⁺, but such conditions led to root growth inhibition.

Excessive uptake of cations over anions across the plasma membrane in root cells leads to acidification of the rhizosphere (Marschner, 1995), explaining the measured proton efflux with K⁺, NH₄⁺ or a combination of both ions in the media (Fig. 7C). The highest H⁺ efflux was consequently measured in the presence of both NH₄⁺ and K⁺, with a reduction upon the co-presence of TEA⁺ and Gd³⁺, in line with the reduced uptake of NH₄⁺ and K⁺. As compared with the H⁺ fluxes measured upon the presence of NH₄⁺, K⁺, and TEA⁺, an unexpected large net H⁺ efflux was observed in the additional presence of Gd³⁺, a situation in which the fluxes of NH₄⁺ and K⁺ were reduced. Gd³⁺ has been shown to inhibit H⁺ influx in the alga Chara corallina (McConneaughy and Falk, 1991) although the precise mechanisms behind this inhibition are unknown. A similar mechanism could lead to the external accumulation of H⁺ observed in this study.

Competition of NH₄⁺ and K⁺ at the site of transporters

In Arabidopsis roots, AKT1 and particularly HAK5 are the main players in high affinity K⁺ uptake (Rubio et al., 2008). AKT1 transports K⁺ over a broad range of concentrations and is suggested to facilitate NH₄⁺-insensitive K⁺ uptake (Hirsch et al., 1998; Angeles Martínez-Cordero et al., 2005). However, a study using insertion lines showed that transport of K⁺ by both AKT1 and HAK5 is sensitive to NH₄⁺ (Rubio et al., 2008). Our results from heterologous expression of AKT1 and HAK5 in the yeast mutant Δtrk1,2 support this earlier study, although inhibition of K⁺ transport by NH₄⁺ was more pronounced in yeast expressing HAK5 relative to AKT1. In the presence of NH₄⁺, induction of HAK5 is suppressed by low K⁺ concentration (Qi et al., 2008; Rubio et al., 2008), suggesting that AKT1 is the main transporter for high affinity K⁺ uptake in the presence of NH₄⁺ (Rubio et al., 2008) and that HAK5 does not play a role in the inhibition of K⁺ uptake by NH₄⁺ in Arabidopsis. However, in the tomato variety Micro Tom, the gene encoding LeHAK5 is highly expressed independent of whether NH₄⁺ is present or not (Nieves-Cordones et al., 2007).

HKT isoforms in various species have been shown to function as K⁺/Na⁺- or Na⁺/Na⁺ co-transporters (Rodríguez-Navarro and Rubio, 2006). Our data suggest that HvHKT2;1 and AtAKT1, when expressed in yeast, are able to transport NH₄⁺ and K⁺. However, the barley cDNA library screen in yeast only retrieved HvHKT2;1 clones, although transformants were grown on various concentrations of K⁺ and NH₄⁺. Other isoforms may not have been functionally expressed in our yeast heterologous expression system as was also the case for AtHAK5 in yeast (Rubio et al., 2000). Lack of functional expression may similarly explain results from a wheat library screen in yeast for the complementation of potassium transport deficiency, which only retrieved TaHKT2;1 and TaLCT1 (Schachtman et al., 1997).

The library screen resulted in the identification of both full-length and truncated versions of HvHKT2;1 with no apparent correlation as to the particular concentrations of NH₄⁺ and K⁺ and the pH of the media. However, the closer characterization on media with lower galactose concentration and without succinic acid as buffer revealed profound differences between full-length HKT2;1 and ΔN₆₂-hkt2;1. Full length HvHKT2;1 best supported growth of Δmep1-3 and Δmep1-3 Δtrk1,2, whereas ΔN₆₂-hkt2;1 complemented better in Δtrk1,2, i.e. in the presence of MEP1-3. This may suggest a cross-regulation between HKT2;1 and Mep/AMT homologues that deserves further investigation. Haro et al. (2005) reported differences in the transport of full-length versus truncated versions of HvHKT2;1. A recent investigation by the same group suggested that it is the expression level of HKT2;1, rather than its truncation, that is responsible for variation in activity (Bañuelos et al., 2008). These differences in expression levels were due to small ORFs in artificial UTRs resulting from the plasmid. We do not know how truncation in our experiments affected the expression level of the proteins. However, we did not observe ORFs in the polylinker of our plasmid, suggesting that the level of expression of both HvHKT2;1 and ΔN₆₂-hkt2;1 was similar.

How could NH₄⁺ uptake through K⁺ transport proteins and NSCCs result in NH₄⁺ toxicity?

Recently, it was suggested that growth depression of barley plants at high NH₄⁺ was due to a futile cycling, involving uptake of NH₄⁺ followed by active extrusion at the plasma membrane (Britto et al., 2001). While such a mechanism can well explain growth depression, it does not explain why plants would aim to eliminate excess NH₄⁺. Extrusion of NH₄⁺ should be the cellular response to avoid the accumulation of NH₄⁺, which is critical at high concentration. Here, the cost of extrusion would represent a secondary but, potentially, rather serious problem. Alternatively, the cycling of NH₄⁺ itself would be an unavoidable process. In such a scenario, futile cycling could indeed be the cause of NH₄⁺ toxicity. Yet, such cycling would require uptake and extrusion mechanisms.

It is well accepted that NH₄⁺ transporters of the AMT family are highly specific for NH₄⁺ over K⁺ (Ninnemann et al., 1994). Recent studies on the structure of AmtB have suggested that discrimination of K⁺ against NH₄⁺ in AMT homologues is due to a de-protonation process of NH₄⁺ in the transporter (Khademi et al., 2004). The E. coli AmtB binds NH₄⁺ with high affinity and, following de-protonation, NH₃, and possibly H⁺, is transported through the channel. Thus, AmtB could be considered a catalytic channel that modifies its substrate with the aim of discriminating between NH₄⁺ and the highly similar K⁺ ion. Although it is a matter of controversy if the transport via AmtB is electrogenic (Javelle et al., 2008), de-protonation at S219 appears to be a prerequisite for transport (Ishikita and Knapp, 2007) and may serve a function to discriminate between NH₄⁺ and K⁺. Similar roles may apply to plant AMTs which have been shown to perform electrogenic transport (Ludewig et al., 2002). Ion transporters and channels transporting K⁺ do not possess such a mechanism of ion discrimination, providing an explanation as to why K⁺ transporters and NSCCs do take up NH₄⁺ even when NH₄⁺ uptake by AMTs is down-regulated. The finding that K⁺ transporters and channels as well as NSCCs transport NH₄⁺ provides a plausible explanation for the uptake of NH₄⁺, at least at low external K⁺, being an unavoidable process due to the physico-chemical similarities of both ions.

In plants, plasma membrane transporters for active extrusion of NH₄⁺ as postulated by Britto et al. (2001) have not been identified. An alternative explanation for an active, energy-consuming extrusion mechanism could be passive diffusion of ammonia (NH₃) out of the cytoplasm either into the vacuole or across the plasma membrane into the apoplast. The vacuole is likely to be the primary compartment trapping NH₄⁺ due to the presence of NH₃ permeable tonoplast intrinsic proteins (Jahn et al., 2004; Loque et al., 2005). Such a role for TIPs in an acid trap mechanism for NH₄⁺ has recently been suggested (Ludewig et al., 2007). The vacuolar compartment, however, can only provide a limited volume for buffering of excess cytoplasmic NH₄⁺. Once this buffer capacity is exhausted, excess NH₄⁺ continuously provided from the external medium could diffuse in the form of uncharged NH₃ along its concentration gradient back across the plasma membrane and into the apoplast. The concentration gradient for NH₃ is maintained by the pH gradient across the plasma membrane. Thus, the continuous uptake of NH₄⁺ into the cell versus diffusion of NH₃ out of the cell would result in a futile cycling of H⁺ across the plasma membrane, and part of this cycle would be catalysed by the ATP-fuelled plasma membrane H⁺-ATPase. Obtaining unambiguous proof for such an hypothesis is challenging, as it is difficult to demonstrate the form in which NH₄⁺/NH₃ is transported in and out of the cell. The observation that the addition of TEA⁺ inhibited NH₄⁺-induced H⁺ efflux from Arabidopsis roots (Fig. 7) however, supports the interpretation that the activity of the plasma membrane H⁺-ATPase represents the energy-consuming step rather than the NH₄⁺ extrusion.

Glossary

Abbreviations

FW

fresh weight

ICP-OES

Inductively Coupled Plasma-Optical Emission Spectrometry

MIFE

Microelectrode Ion Flux Estimation

NSCC

non-selective cation channel

ORF

open reading frame

TEA

tetraethyl ammonium

UTR

untranslated region

YNB

yeast nitrogen base