JXB Advance Access originally published online on May 7, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Experimental Botany, Vol. 55, No. 401, pp. 1293-1305, June 1, 2004
© 2004 Oxford University Press
FOCUS PAPER |
Regulatory levels for the transport of ammonium in plant roots
Received 26 January 2004; Accepted 15 March 2004
1Institut für Pflanzenernährung, Universität Hohenheim, D-70593 Stuttgart, Germany
* To whom correspondence should be addressed. Fax: +49 711 459 3295. E-mail: vonwiren{at}uni-hohenheim.de
 |
Abstract |
|---|
Top
Abstract
IntroductionAmmonium is an attractive nitrogen form for root uptake due to its permanent availability and the reduced state of the nitrogen. On the other hand, ammonium fluxes are difficult to control because ammonium represents an equilibrium between NH+4 and NH3, which are two N forms with different membrane permeabilities. There is increasing evidence that AMT-type ammonium transporters represent the major entry pathways for root uptake of NH+4. Since excess uptake of ammonium might cause toxicity and since ammonium is also released from catabolic processes within the cell, ammonium uptake across the root plasma membrane has to be tightly regulated. To take over a function in cellular ammonium homeostasis, various AMT transporters are synthesized that differ in their biochemical properties, their localization, and in their regulation at the transcriptional level. At the same time, AMT-driven transport is subject to control by the nitrogen status of a local root portion as well as of the whole plant. In this review, the focus is on the different levels at which AMT-dependent ammonium uptake is regulated and the gaps in current knowledge are highlighted.
Key words: Ammonium, AMT, plant, regulation, transport, uptake.
|
Introduction |
|---|
The ability of roots for nutrient uptake and translocation at a rate that matches the demand of the different plant organs is fundamental for efficient plant development and reproduction. If nutrient availability declines or abiotic and biotic stresses inhibit proper root function, plants run into nutrient deficiencies that decrease their ecological competitiveness or agricultural productivity (Marschner, 1995). Since the sizes and locations of plant-available nutrient pools in soils are often highly variable and conditions for nutrient uptake, which are influenced by drought, oxygen depletion etc., are rarely optimal during a whole vegetation period, transient nutrient deficiencies must be expected to occur frequently. This is of particular importance for nitrogen (N), which is the mineral nutrient required in the largest amounts and is one of the most limiting factors in agricultural plant production.
To direct morphological and physiological responses to nutrient deficiencies, plants require, on the one hand, nutrient transport systems with high flexibility regarding internal nutrient demand, external substrate availability, and the spatial distribution of nutrient sources within the exploitable soil volume. On the other hand, plants require sensing systems scanning the external substrate concentration in the rooted area and signalling to the plant in which direction a further development of the root system could be most promising. For a few nutrients, like nitrate or phosphate, lateral root formation has been shown to respond to local nutrient availability (Drew, 1975), and in the case of nitrate, the putative transcription factor, ANR1, has been proposed as an essential constituent of the signalling cascade between localized nitrate availability and lateral root formation in Arabidopsis (Zhang and Forde, 1998).
Nutrient-sensing systems in roots could either measure external nutrient concentrations and translate them into transferable signals or measure the internal nutrient concentration in distinct root cells. This could preferably occur in root hairs that are spatially most exposed to sense nutrients and that vary length and density in response to external nutrient supply (Ma et al., 2001; Schmidt and Schikora, 2001). In yeast, the outgrowth of pseudohyphae, which are expanded cell types that are well adapted to foraging a substrate deeply, is a response to low ammonium availability and is triggered by the high-affinity ammonium transporter Mep2 (Lorenz and Heitmann, 1998). While plant membrane proteins involved in sensing external nutrient availability still await to be uncovered, internal sensing systems for Fe or nitrogenous compounds have been proposed on the basis of mutant analysis or the behaviour of transgenic plants (Coruzzi and Bush, 2001; Schmidt, 2003). A tight interplay between sensor proteins and highly flexible transport systems is expected to be of particular importance for the co-ordination of N uptake, since the heterogenic turnover of organic matter in soils usually produces a high spatial variability of N pools.
As the primary inorganic nitrogenous compound released from mineralized organic matter, ammonium plays, besides its nitrification product nitrate, a key role in N nutrition for plants and soil microorganisms. This is emphasized by its preferential uptake in many plants when supplied in equimolar concentrations with nitrate (Sasakawa and Yamamoto, 1978; Gazzarrini et al., 1999) and particularly holds true as long as ammonium supply is low. By contrast, at higher and exclusive supply, ammonium tends to inhibit plant growth (Walch-Liu et al., 2001) or to generate toxicity (Britto and Kronzucker, 2002). To adjust cellular ammonium levels that vary not only in response to the uptake of external ammonium but also to intracellular amino acid catabolism, ammonium transport processes need to be tightly regulated. The present review attempts to focus on the different levels at which ammonium transport processes are controlled in plants. Regulatory levels of ammonium transport processes are considered from the level of individual transport proteins to the level of the whole plant. For these different regulatory levels that control ammonium uptake, present concepts are outlined and open questions as well as research needs are highlighted.
|
NH+4 as the dominant species for controlled membrane transport
|
|---|
Early electrophysiological studies in the unicellular algae Chara australis showed that the addition of ammonium or methylammonium into the external medium provoked positive inward currents across the plasma membrane (Walker et al., 1979a). These methyl-ammonium-induced currents further increased with more negative membrane potentials and were accompanied by a depolarization of the membrane potential, indicating membrane transport of a charged species. Increasing the pH of the medium above neutral pH decreased inward currents, which was best explained by the concomitant decrease in the ratio of the charged species versus the free base (Walker et al., 1979b). Likewise, in root cells of rice or Arabidopsis plants ammonium-induced membrane depolarization increased in a nearly saturable manner at micromolar concentrations and was also affected by the N nutritional status of the plants (Wang et al., 1994; Shelden et al., 2001). Thus, electrophysiological studies provided strong evidence that charged NH+4 must be the major substrate for regulated root uptake of ammonium.
On the other hand, uptake studies with 14C-labelled methylammonium in Chara cells showed that the substrate was efficiently absorbed over a wide pH range, although at high pH ammonium uptake was no longer closely related with a depolarizing effect on the plasma membrane potential (Smith and Walker, 1978). This observation agreed with the view of an increasing permeability of the uncharged species at high pH (Walker et al., 1979b) and pointed to the parallel existence of another transport pathway, that is specialized for the transport of the uncharged form.
Using stopped-flow spectrofluorimetry, the NH3 permeability of peribacteroid membranes from soybean nodules was tested, for which a high ammonium permeability had been reported previously (Tyerman et al., 1995). It was found that permeation of NH3 was inhibited by mercury, similar to that of water. Since NH3 permeation through protein-free liposomes remained unaffected by mercury, it was concluded that a part of the transmembrane passage of NH3 across vesicle membranes was mediated by proteins, which showed a similar feature like water transport via aquaporins (Niemietz and Tyerman, 2000).
Nakhoul et al. (2001) expressed the human aquaporin AQP1 in Xenopus oocytes and found that the internal pH in the oocyte increased with increasing pH in the outer solution when ammonium was supplied. Since this internal pH shift was not associated with a change in membrane potential, it was concluded that AQP1 enhanced NH3 permeability. However, due to spatial heterogenity in oocytes, the measurement of the internal pH might be complicated and other studies were not able to verify NH3 transport by AQP1 (Yang et al., 2000; Verkman, 2002). Thus, future approaches should attempt direct measurements of NH3 transport and to reconstitute purified aquaporins in artificial bilayers, employ mutant aquaporins in heterologous expression systems to dissect water from NH3 transport, or demonstrate NH3-related phenotypes in transgenic plants with altered expression or activity of aquaporins.
Soupene et al. (1998, 2001, 2002) proposed NH3 as a primary substrate for bacterial and fungal ammonium transporters of the AMT/MEP family as well. This proposition was based on the observation that, at micromolar concentrations of ammonium, growth of AMT-defective E. coli was impaired at low pH, but not at neutral pH. Functional analysis by two-electrode voltage clamp of two tomato AMTs expressed in Xenopus oocytes, however, showed that ammonium-induced positive inward currents did not change significantly over a pH range of 5.5–8.5. Since NH3 alone could not account for the observed currents and protons did not appear as a co-substrate, NH+4 uniport is the most likely transport mechanism for AMT1-type transporters from plants (Ludewig et al., 2002, 2003).
In intact root cells or in protoplasts, ammonium produced similar, but weaker, currents compared with potassium (Gassmann and Schroeder, 1994; Demidchik and Tester, 2002). Since a single amino acid substitution in a potassium channel can dramatically increase ammonium permeability (Uozumi et al., 1995), it is likely that ammonium inhibition of potassium transport by channels like AKT2 (Dennison et al., 2001) is a result of direct substrate competition (White, 1996). Whether this ammonium transport through potassium or non-selective cation channels also contributes to ammonium nutrition, remains to be demonstrated. Until then, ammonium transport may be considered as a side activity of potassium and non-selective cation channels.
|
The AMT family of ammonium transporters in plants |
|---|
After the first ammonium transporter AtAMT1;1 was isolated from the model plant Arabidopsis thaliana by growth complementation of a yeast mutant defective in ammonium uptake (Ninnemann et al., 1994), sequencing of the Arabidopsis genome allowed another five homologous sequences (Gazzarrini et al., 1999; Sohlenkamp et al., 2000) to be discovered. Phylogenetic analysis revealed that AtAMT1;1 to AtAMT1;5 showed the highest homology to each other and, clustered with cyanobacterial ammonium transporters, forming the AMT subfamily of ammonium transporters. AtAMT2 was sequence-wise more closely related to the ammonium transporters Mep1, 2 and 3 from Saccharomyces cerevisiae and with AmtB from E. coli, forming, together with other prokaryotic homologues, the MEP subfamily (Ludewig et al., 2001). A third set of homologous sequences does not encompass any plant protein so far, but mainly includes human and animal Rhesus blood group antigens, thus forming the Rh subfamily of ammonium transporters.
Screening of a database with >150 000 ESTs from tomato, of cDNA libraries from different tissues, and of the tomato genome using a PCR-based approach allowed only three ammonium transporters of the AMT1 family to be isolated repeatedly (Lauter et al., 1996; von Wirén et al., 2000) and a single putative orthologue of AtAMT2, LeAMT2 (BI922102 [GenBank] ).
With the genome sequencing of the graminaceous model plant rice, a plant species had been chosen that is, unlike Arabidopsis and tomato, highly adapted to ammonium-based nitrogen nutrition (Yoshida, 1981). Interestingly, rice possesses a much higher number of ammonium transporter genes than Arabidopis and tomato (Suenaga et al., 2003; Sonoda et al., 2003a), while, with regard to nitrate transporters, their number appears not to be higher in rice compared with Arabidopsis and tomato. Taken with caution, a comparison of the number of AMT genes in different species might suggest that plant species from different environments organize ammonium transport with a fairly unequal number of ammonium transporters. However, expression analysis and functionality tests of all these transporters are required before firm conclusions can be drawn.
A phylogenetic analysis for AMT-homologous sequences from Arabidopsis, tomato, and rice indicated that rice AMTs subdivide into four clades (Suenaga et al., 2003), while those of Arabidopsis divide into two clades only (Ludewig et al., 2001). Using another type of phylogenetic analysis that includes all available AMT sequences from Arabidopsis, tomato, and rice, except LeAMT2 for which only a short amino acid sequence is available, indicates that the rice sequences also generally cluster into two clades, although OsAMT3;2 and OsAMT4;1 were set apart from OsAMT2-related sequences or from the OsAMT3;1 and 3;3 branch (Fig. 1). The relatively large sequence distance between the AMT1 and AMT2/MEP subfamilies might point to a different selection pressure under which these transporters have evolved. It will thus be of interest to uncover whether the observed phylogenetic distance between these two subfamilies also translates into different transport functions.
|
Hydrophobicity plots predicted that AMT polypeptides represent integral membrane proteins. A summary of several actual computer analyses predicted a protein topology for plant AMTs of 11 transmembrane-spanning domains with an extracytosolic N-terminus and a cytosolic C-terminus (www.aramemnon.de; Schwacke et al., 2003). These predictions were supported by a topological analysis of the yeast Mep2 protein, which carries an N-glycosylation site at its extracytosolic N-terminus (Marini and André, 2000). Unlike the plant and yeast homologues, the E. coli AmtB protein contains an additional, twelfth transmembrane spanning domain at its N-terminus (Thomas et al., 2000) that is cleaved off in the mature protein (D Blakey, M Merrick, unpublished results). Although computer simulation also predicted with considerable likelihood the existence of N-terminal signal peptides for chloroplastic localization (TargetP, www.cbs.dtu.dk) in AtAMT1;1, AtAMT1;2, LeAMT1;1, and LeAMT1;3, chloroplast import assays showed that in vitro translated radio-labelled proteins were not incorporated into isolated chloroplasts (Shelden et al., 2001; U Flügge, N von Wirén, unpublished results).
From the computational analysis of AMT protein sequences, it is difficult to identify protein domains required for the ammonium transport function (reviewed in von Wirén and Merrick, 2004). To date, understanding of the structure–function relationships relies mostly on the functional analysis of mutated AMT proteins in heterologous systems. Individual point mutations have been reported that cause single amino acid exchanges in AMT sequences with a subsequent loss of functionality (Salvemini et al., 2001; Monahan et al., 2002). Deriving conclusions from mutational analysis, however, requires verification as to whether the mutant proteins are expressed and properly transported to their target membrane. For plant AMTs this requirement has only been satisfactorily addressed in the case of LeAMT1;1, in which an amino acid exchange in the C-terminus (G458D) caused loss of function while protein synthesis and targeting were apparently not affected (Ludewig et al., 2003).
|
Regulation of NH+4 transport by AMTs at the mRNA level |
|---|
Uptake studies with 15N-labelled ammonium in plant roots showed a continuous increase of ammonium influx from the start until the end of the light period, followed by a sharp decline after the light was turned off (Ourry et al., 1996; Gazzarrini et al., 1999). These diurnal changes in ammonium uptake capacity correlated with AMT1 transcript levels in roots, in particular for AtAMT1;3 in Arabidopsis. Since sucrose supply during the dark period could prevent a drop down of AtAMT1;1, 1;2, and 1;3 mRNA levels and maintain ammonium influx at a similar rate as in the light (Lejay et al., 2003), it was suggested that photoassimilates regulate ammonium transport.
Gene induction by sugars is not specific to AMTs. Similar results have been observed with transporters for nitrate and other mineral elements, that all showed a diurnal regulation at the transcript level (Lejay et al., 2003). In the case of the sugar response of AtNRT2;1, a gene encoding a high-affinity nitrate transporter from Arabidopsis, an analysis of Arabidopsis mutants, that are defective in various steps of the putative sugar signalling pathway, and of the effect of non-metabolizable sugar analogues on gene expression suggested that a downstream product of glucose from the glycolytic pathway should trigger the transcriptional regulation of AtNRT2;1 (Lejay et al., 2003). Future studies on a glycolysis-dependent regulation of AMTs should also consider the analysis of metabolic pools in mutant lines. Perhaps the screening for loss of sugar inducibility of EMS-mutagenized Arabidopsis lines expressing reporter genes controlled by diurnally regulated AMT-promoters may help to identify genes that control individual steps in this signalling pathway.
Transcript levels of AMT genes strongly respond to the N nutritional status of the plant. Within 3 d of N deficiency mRNA levels of AtAMT1;1 and 1;3 in Arabidopsis roots clearly increased, while AtAMT1;2 and AtAMT2;1 tended to increase after a more extended period of N deficiency (Gazzarrini et al., 1999; Sohlenkamp et al., 2000; D Loqué, A Gojon, N von Wirén, unpublished results). The steep increase in AtAMT1;1 mRNA suggested that the gene product makes a major contribution to the overall ammonium uptake capacity in roots. Against this expectation, it was surprising that an atamt1;1::T-DNA (atamt1;1-1) insertion line showed only a 30% decrease of high-affinity ammonium uptake (Kaiser et al., 2002). Thus, at least partially overlapping functions should be expected among the different AMT homologues. While Kaiser et al. (2002) found, by means of RT-PCR, a higher expression of AtAMT1;3 and AtAMT2;1 in N-deficient roots of the atamt1;1-1 insertion line, another study could not find any significant differences in AMT gene expression between the wt and the same insertion line (Fig. 2). A compensatory response of other AMTs to the loss of AtAMT1;1 expression thus remains to be substantiated. Ongoing studies should also consider using western blot studies to investigate a possible compensation at the AMT protein level, besides the analysis of a recomplemented atamt1;1-1, insertion line.
|
As for diurnal regulation, the mechanism regulating ammonium transport by AtAMTs under N-limiting conditions remains unknown. Gansel et al. (2001) suggested that ammonium uptake is predominantly regulated by the local N status of the roots rather than by the N status of the whole plant as in the case of nitrate uptake. Moreover, evidence was raised that ammonium uptake is regulated at the transcriptional and post-transcriptional levels simultaneously (Rawat et al., 1999). On the one hand, elevated ammonium concentrations in the cytoplasm inversely correlated with ammonium influx before transcriptional changes in AMT mRNA levels were observed. On the other hand, AtAMT1;1 gene expression negatively correlated with root glutamine concentrations, which was indicative for glutamine being the metabolic trigger down-regulating the transcription of AtAMT1;1 (Rawat et al., 1999).
Regulation of gene expression among AMT homologues can also behave in an opposite way and increase under adequate or high N supply. An up-regulation after resupply of ammonium was first observed for LeAMT1;2 mRNA levels in tomato roots and then for OsAMT1;1 and 1;2 in rice roots (von Wirén et al., 2000; Sonoda et al., 2003a). In rice, ammonium could be replaced by glutamine that would still trigger the induction of OsAMT1;1 and 1;2 (Sonoda et al., 2003b). Thus, glutamine might act as a metabolic trigger for the up- and down-regulation of AMT genes depending on the individual AMT gene and plant species (Rawat et al., 1999; Sonoda et al., 2003b). Moreover, in tomato, LeAMT1;2 was not exclusively induced after the resupply of ammonium but was also induced after the resupply of nitrate (Lauter et al., 1996; Wang et al., 2001). This regulatory feature might point to an additional function of LeAMT1;2 in ammonium retrieval, since nitrate supply also leads to an efflux of ammonium (Morgan and Jackson, 1988). The existence of N-inducible AMTs tempts to speculate that these AMTs may have a higher transport capacity compared with their N-derepressed counterparts. An answer to this question, however, awaits the functional analysis of transgenic lines with repressed AMT gene expression levels. The complexity of regulation revealed in the above studies is summarized in Table 1.
|
|
Regulation of AMT-mediated NH+4 transport at the protein level |
|---|
Biochemical properties
To compare kinetic properties among AMT1 transporters, AtAMT1;1, 1;2, and 1;3 were functionally expressed in the yeast triple Mep deletion strain 31019b (
mep1, mep2, mep3), and concentration-dependent inhibition of 14C-methylammonium uptake by NH+4 was determined. Km values obtained for AtAMT1;2 and 1;3 were between 25 and 40 µM for both substrates, while AtAMT1;1 seemed to discriminate greatly between both substrates with Km values of 8 µM for methylammonium but <0.5 µM for NH+4 (Gazzarrini et al., 1999). Since AtAMT1;1 responded to N deficiency with the most dramatic increase in mRNA levels among all AMT genes, it was concluded that up-regulation of AtAMT1;1 under N starvation might be evolutionarily linked with a high substrate affinity of the corresponding gene product. From a biochemical perspective, differing substrate affinities of individual transporters might then represent one factor for the existence of multiple AMT1 homologues (Gazzarrini et al., 1999). In a more direct approach, concentration- and pH-dependent 13N-ammonium uptake was determined in AtAMT1;1- and AtAMT2-expressing yeast cells. For both transporters a Km value of approximately 20 µM was obtained that was hardly affected by pH in a range of 5.0 and 7.5 (Sohlenkamp et al., 2002). An overview on substrate affinities determined for plant AMT proteins is summarized in Table 1.
The substrate affinity of AtAMT1;1 was further investigated in the T-DNA insertion line atamt1;1-1, that showed a complete loss of AtAMT1;1 mRNA in Arabidopsis roots (Kaiser et al., 2002). Between 10 and 250 µM 13N-labelled ammonium, influx in the insertion line was approximately 30% lower compared with the wild type when plants were precultured under N deficiency. This observation corresponded with the reduction of transport capacity with an approximate Km value between 10 and 20 µM. Although these two experimental approaches yielded independently a similar substrate affinity for AtAMT1;1, it should be taken into consideration that a second lower Km value in the nanomolar range might have been missed, since uptake experiments were conducted at concentrations above 10 µM external ammonium (Kaiser et al., 2002; Sohlenkamp et al., 2002).
In atamt1;1-1, Kaiser et al. (2002) observed that low-affinity uptake of 15N-ammonium, as determined in the concentration range between 0.25 and 10 mM, was even higher than in the wild type, which was explained by a compensatory transcriptional up-regulation of other AMTs in the amt1;1-1 insertion line. With regard to possible interactions between AMTs (Ludewig et al., 2003) compensation for the loss of an individual AMT transporter, however, can also occur at the protein level. It is therefore essential to verify Km values of AMTs as obtained in heterologous systems in transgenic or mutant plants with low ammonium uptake activity. In view of the, at least partially, overlapping transport capacities mediated by AtAMTs, this will require the generation of Arabidopsis lines with multiple insertions in AMT genes, that are then retransformed with individual AMTs for subsequent functional analysis.
Alternatively, the substrate affinity of AMTs was studied by two-electrode voltage clamp in Xenopus oocytes expressing tomato AMTs. While LeAMT1;3 could not be functionally expressed in oocytes so far, LeAMT1;1 turned out to show a fairly low Km value of approximately 10 µM and LeAMT1;2 of approximately 60 µM (Ludewig et al., 2002, 2003). Thus, the Km of LeAMT1;1 closely matched that of intact tomato roots, which saturated with a Km of approximately 8.5 mM when cultivated under constant supply of 50 µM ammonium (Smart and Bloom, 1988).
A comparison of substrate affinities for ammonium transporters of the MEP family in yeast yielded micromolar Km values for Mep1 and Mep2, but a millimolar Km value for Mep3, indicating that one transporter subfamily can encompass high- and low-affinity transporters at the same time (Marini et al., 1997). In an attempt to verify Km values for AtAMT1;1 and AtAMT1;2, Shelden et al. (2001) employed AMT-transformed yeast cells for uptake studies using 14C-labelled methylammonium. Besides the confirmation of high-affinity transport activity, they reported an additional uptake capacity of AtAMT1;2 at millimolar methylammonium concentrations, which would be indicative for a dual affinity of AtAMT1;2. Measuring low-affinity methyl-ammonium uptake in the yeast triple Mep deletion strain, however, appeared to be problematic, since a considerable amount of substrate enters this strain via so-far-unknown low-affinity transporters. This residual background of ammonium uptake in the triple mep mutant prompted Marini et al. (1997) to integrate uptake periods over a longer period of time and to estimate the substrate affinity of the low-affinity Mep3 via ammonium depletion from the external medium. Thus, before a dual Km for AtAMT1;2 can be proposed, more detailed transport studies are required, such as those that have been conducted, for example, in case of the dual-affinity transporters AtNRT1;1 and AtKUP1 (Liu and Tsay, 2003; Fu and Luan, 1998).
Investigations on the substrate selectivity of AMT transporters generally revealed a high capability of AMT proteins to discriminate ammonium from other potential substrates. AMT1 transporters hardly transported any potassium or other monovalent cations with the exception of methylammonium (Ninnemann et al., 1994; Ludewig et al., 2002, 2003). In LeAMT1;1- and LeAMT1;2-expressing oocytes the preferential transport of ammonium over methylammonium was expressed in an approximately 5-fold and more than 10-fold weaker current, respectively. A similar difference in the uptake ratio between both substrates was observed in intact roots of tomato (Kosola and Bloom, 1994).
Protein–protein interactions
The first experimental evidence for the possible regulation of ammonium transporter activity at the protein level derived from yeast. The Saccharomyces cerevisiae yeast genome encodes three AMT homologues (MEP1-3), which all mediate ammonium uptake when expressed individually in a triple Mep deletion strain (Marini et al., 1997). However, Mep3 was non-functional in a mep1-1 mep-2 background (Dubois and Grenson, 1979), that carries a complete deletion of MEP2 and a point mutation in MEP1 leading to a single amino acid exchange in the C-tail of Mep1 (G413D) (Marini et al., 2000b). This point mutation completely inactivated Mep1 and trans-inhibited the activity of Mep3 resulting in an ammonium uptake-defective phenotype, which is similar to the triple mep deletion strain 31019b. Highly similar results were obtained by introducing the corresponding amino acid exchange in the ammonium transporter protein MepA from Aspergillus (Monahan et al., 2002). These results suggested that individual ammonium transporters of the Mep subfamily might interact via their cytosolic C-termini.
The equivalent amino acid exchange in LeAMT1;1 from tomato (G458D) also caused loss of function (Ludewig et al., 2003). Since a G458D-GFP fusion protein still localized to the plasma membrane when expressed in yeast, an effect of the amino acid substitution on protein targeting was excluded. Voltage-clamp analysis of the G458D mutant protein in Xenopus oocytes confirmed the essentiality of the conserved glycine residue for functional ammonium transport. To investigate a possible interaction between AMT1 transporters, the LeAMT1;1 wild-type and G458D mutant proteins were co-expressed in oocytes, which resulted in a dramatic decrease of LeAMT1;1 wild-type activity, while an amino acid permease co-expressed with LeAMT1;1G458D did not lose activity (Ludewig et al., 2003). Since LeAMT1;1G458D also inhibited ammonium transport by a co-expressed LeAMT1;2 wild-type protein and, since, in turn, an exchange of the conserved glycine residue in LeAMT1;2 (G465D) inhibited ammonium transport by wild-type LeAMT1;1, it was concluded that LeAMT1;1 and LeAMT1;2 interact, at least when expressed in oocytes. A physical interaction between LeAMT1;1 proteins was further supported using a split-ubiquitin system, which led to the hypothesis that LeAMT1 proteins might form oligomers. Indeed, western blot analysis of microsomal membrane proteins using a specific antibody against LeAMT1;1 detected LeAMT1;1 at a 3–4-fold higher molecular weight than expected for a LeAMT1;1 monomer. Since denaturation of the protein samples prior to SDS-PAGE resulted in a drop of the protein band with high molecular weight to its putative monomeric size, it was concluded that LeAMT1;1 proteins also oligomerize in vivo (Ludewig et al., 2003). Further investigations are required to verify whether homo- and/or hetero-oligomerization occur in vivo and whether unknown proteins interact with LeAMTs and, thereby, contribute to a shift in the apparent protein size. So far, these observations show that LeAMTs can oligomerize and raise the question whether this type of protein–protein interactions might be a regulatory component in AMT protein activity.
|
Regulation of NH+4 uptake at the cellular level |
|---|
Cytoplasmic ammonium pools are not only replenished by ammonium uptake across the plasma membrane, but also by ammonium release from the internal phenylpropanoid metabolism or amino acid catabolism. The latter process is of particular importance in illuminated leaf cells, where ammonium is released during photorespiration from the mitochondria (Leegood et al., 1996) or in senescent tissues, when proteins are broken down to amino acids (Mattsson and Schjoerring, 2003). Since low ammonium concentrations in the cytoplasm have long been regarded as a prerequisite to protect the cell from ammonium toxicity (Marschner, 1995), ammonium-induced processes are to be expected that compensate for excess ammonium generation. Otherwise, cytoplasmic ammonium toxicity might lead to rapid necrosis of the plant tissue. Besides controlling ammonium uptake across the plasma membrane, plant cells can modulate ammonium efflux out of the cytoplasm, ammonium compartmentation into the vacuole, or ammonium assimilation in the cytoplasm or the plastids.
Regulation of cytoplasmic ammonium concentrations by ammonium efflux
Intracellular ammonium concentrations have frequently been estimated on the basis of compartmental efflux analysis or nuclear magnetic resonance studies. Estimates from cytoplasmic ammonium concentrations ranged from 0.01–76 mM (Miller et al., 2001), however, most studies agreed with concentrations ranging from one to several tens of millimolar (reviewed in Britto and Kronzucker, 2002). More recently, measurements with novel ammonium-selective microelectrodes in N-supplied Chara cells also found cytoplasmic ammonium concentrations in the range of 1–10 mM (Wells and Miller, 2000). Since apoplastic ammonium concentrations, that have mostly been determined in leaf cells, seem to be highly buffered around 1 mM (Nielsen and Schjoerring, 1998; Hanstein and Felle, 1999), an outward-oriented ammonium gradient must exist, which is even larger for NH3 due to the high cytoplasmic pH. In the case of NH+4, however, the concentration gradient is compensated for by the electrical gradient allowing secondary active NH+4 accumulation in the cytoplasm (reviewed in Britto et al., 2001a). Due to a less negative membrane potential in root cells of rice compared with barley, rice accumulated a relatively lower amount at high external ammonium supply, which finally led to a higher tolerance against ammonium toxicity (Britto et al., 2001b). Since ammonium accumulation and efflux from barley root cells were larger and coincided with a higher root respiration rate, ammonium efflux was considered to be an energy-demanding process. It was therefore concluded that ammonium efflux and the plasma membrane potential in root cells are crucial factors for tolerance against ammonium toxicity (Britto et al., 2001b). Moreover, the modulating effect of nitrate on ammonium efflux might further support the view that ammonium efflux is regulated (Britto and Kronzucker, 2001). To date, this study’s view on ammonium efflux processes in plants is completely based on correlative and even coincidental data from physiological studies, while the molecular identity of the underlying transporters mediating ammonium export remains unclear. A first step to describe ammonium export at the molecular level might, therefore, be to test the bidirectionality of ammonium transport by AMT proteins.
Following their electrophysiological behaviour in Xenopus oocytes LeAMT1;1 and LeAMT1;2 mediated NH+4 uniport in dependence of the membrane potential and the concentration gradient of NH+4 (Ludewig et al., 2002). In principle, both LeAMTs might also mediate NH+4 efflux, if the electrochemical gradient is reversed. While experimental evidence for an efflux function of plant AMTs is not yet available, supportive evidence comes from the study of the human Rhesus factors that belong to a different subfamily of ammonium transporters (Ludewig et al., 2001). Yeast cells expressing RhAG or RhGK promoted ammonium release to the medium when internal ammonium generation was enhanced through the supply of arginine as a sole N source (Marini et al., 2000a). A similar observation was made in Salmonella typhimurium, where growth rates of an amtB deletion strain were less inhibited by arginine supply as a sole N source compared with the wild type (Soupene et al., 2002). The effects of an amtB disruption on the utilization of ammonium generated internally, as well as on the use of external ammonium, were indicative of a bidirectional function of AmtB.
When culturing yeast on arginine, ammonium release by the triple mep deletion strain was high, but strongly reduced after individual MEP genes were reintegrated (Marini et al., 1997). On the one hand, this clearly indicated that Mep proteins function in the retrieval of ammonium leaking out of the cell, and, on the other, that ammonium release must also follow other pathways than those mediated by MEP transporters. Physiological studies in plants indicated that the transport of NH3 contributes to overall ammonium efflux, since higher tissue ammonium concentrations enhanced the release of gaseous NH3 from leaves (Mattsson and Schjoerring, 2002). Therefore, the physiologically relevant question is to what extent can the plant cell control NH3 transport across the (plasma) membrane. If aquaporins function in transmembrane NH3 fluxes, gating of these channels might represent one regulatory tool.
Regulation of cytoplasmic ammonium concentrations by ammonium compartmentation and assimilation
Since vacuoles can occupy more than 90% of the plant cell, they not only possess a large capacity for ammonium storage but they are also accessible for more reliable ammonium measurements. Vacuolar ammonium concentrations in non-stressed plants ranged between 2 and 45 mM (summarized in Miller et al., 2001), clearly supporting a role of vacuoles as a cellular storage compartment for ammonium. Given an at least two times higher ammonium concentration in the vacuole relative to the cytoplasm, as, for example, observed in Chara cells (Wells and Miller, 2000), and a slightly positive potential across the tonoplast, the electrochemical gradient would drive NH3 import into, but NH+4 export out of, the vacuole into the cytoplasm (Britto et al., 2001a). In the case of NH+4, transport across the tonoplast would require electrogenic ammonium transporters, as, for example, have been described by patch-clamp measurements of mesophyll vacuoles (Brüggemann et al., 1999). These fast-activating cation channels showed a high selectively for monovalent cations, and in particular for NH+4. In addition, the open probability of these channels was increased in the presence of ammonium.
Interestingly, when vacuolar acidification is inhibited, as in the vma mutant of yeast, ammonium supply to the medium promoted NH+4 loading of the vacuole and vacuolar acidification. This supported the notion that, with decreasing tonoplast membrane potential, NH+4 loading of the vacuole becomes favourable and that NH+4 might then act as a protonophore (Plant et al., 1999). Otherwise, vacuolar loading with NH+4 should be an energy-demanding process, while vacuolar loading with NH3 would occur ‘downhill’ and would mainly be driven by the acidic pH in the vacuole. Although the molecular identity of transporters for vacuolar loading and unloading remains speculative, again, aquaporins, as, for example, from the TIP subfamily, might be promising candidates. Individual members of the TIP subfamily have been demonstrated to transport, besides water, other neutral nitrogenous solutes, such as urea also (Liu et al., 2003).
Cytoplasmic concentrations and transmembrane fluxes of ammonium are also dependent on the rate of ammonium assimilation. From inhibition studies with the glutamine synthetase inhibitor, MSX, it remains unclear which metabolic compound leads to down-regulation of ammonium influx in roots. The large number of controversial results might reflect that plant genotype and the actual N nutritional status might influence the type of action of MSX (reviewed in Glass et al., 2002). This is emphasized by the occurrence of ammonium-inducible and ammonium-repressible AMT transporter genes.
In rice, ammonium-inducible transcripts of NADH-GOGAT colocalized with those of ammonium-inducible OsAMT1;2 in root tips (Ishiyama et al. 2003; Sonoda et al., 2003b), where usually the highest ammonium uptake capacity is found (Colmer and Bloom, 1998). Moreover, ammonium-inducible gene expression as analysed by RT-PCR revealed an almost identical time-dependent kinetics (Sonoda et al., 2003b). These observations suggest that ammonium uptake and assimilation are closely synchronized, at least in N-demanding rice roots.
Relative to roots the situation might differ in leaves. If ammonium assimilation was repressed in glutamine synthetase-defective barley mutants, ammonium concentrations in the leaf tissue increased, as did NH3 emission from leaves (Mattsson et al., 1997). By contrast, investigating temperature-dependent photorespiratory ammonium production in GS2-antisense lines from oilseed rape, Husted et al. (2002) showed that photorespiration as a major NH3-releasing metabolic process neither controlled tissue levels of ammonium nor NH3 emission rates. Thus, perhaps additional factors, such as compensatory metabolic processes under GS-repressed conditions, must be considered before the effect of ammonium assimilation and reassimilation on transmembrane fluxes of ammonium can be predicted. Interestingly, in rape, transcript levels of an ammonium-inducible AMT homologue, BnAMT1;2, seemed to be co-regulated with GS2 activity (Pearson et al., 2002). In further studies it will be important to verify whether, in AMT-overexpressing plant lines, NH3 emission can be reduced by a higher capacity for ammonium retrieval from the apoplast.
Although efflux, compartmentation, and assimilation of ammonium certainly play a role in the regulation of the cytoplasmic ammonium pool and thereby might feedback regulate AMT-driven ammonium transport, a direct regulatory link between one of these processes and AMTs has not yet been observed. By contrast, in E. coli AmtB activity seems directly regulated by the PII protein which is thought to sense the carbon and nitrogen status in the cell (Coutts et al., 2002).
|
Regulation of root NH+4 uptake at the whole-plant level |
|---|
A dual regulation has been proposed for the root uptake of nitrate, a systemic regulation by a shoot-to-root signal and a local regulation triggered by the availability of the substrate in a given root zone. In a split-root experiment high-affinity nitrate transport was high in a nitrate-supplied root portion, but low in N-deficient roots, while ammonium influx was repressed in those root portions which were supplied with ammonium (Gansel et al., 2001). It was concluded that only nitrate influx must have been controlled by a shoot-derived signal, while this putative signal was probably too weak to overcome the ammonium repression in the ammonium-supplied root portion (Gansel et al., 2001). A shoot-to-root signal may thus be considered as being of minor importance relative to the effect of the local N status that dominates control over ammonium influx.
A local regulation of ammonium influx is probably dependent on the ammonium assimilation rate in roots. Compared with nitrate, assimilation of ammonium mainly takes place in roots and seems strongly dependent on the cytosolic isoform of glutamine synthetase (Tobin and Yamaya, 2001). Ammonium-grown plants often show a higher concentration of amino acids than nitrate-grown plants (Allen and Smith, 1986; Causin and Barneix, 1993). Thus, ammonium nutrition leads to an immediate increase of the cellular ammonium and amino acid pools and might then not involve a shoot signal for the down-regulation of ammonium influx. This is supported by an enhanced ammonium uptake capacity in Arabidopsis roots after the transfer from ammonium nitrate to nitrate nutrition (Gazzarrini et al., 1999), which was under conditions where the shoots were unlikely to have suffered from N deficiency.
The root-to-shoot translocation of ammonium or of ammonium-derived nitrogen is rapid. Using positron-emitting tracer imaging, 13N from root-applied 13N-labelled ammonium was found in the shoot tissue within less than 2 min. In the presence of MSX, however, root-to-shoot translocation was strongly reduced, which emphasizes the inhibitory effect of local ammonium accumulation on influx (Kiyomiya et al., 2001). In N-deficient rice roots 13N translocation from 13N-labelled ammonium decreased with increasing root demand (Kronzucker et al., 1998) suggesting that the root serves first before surplus ammonium-N is translocated further to the shoot. In both cases, however, it remained unclear to what extent ammonium-derived 13N was converted into amino acids before translocation.
In the case of nitrate, the nature of a shoot-derived signal was initially proposed to consist of a fraction of the amino acids translocated in the phloem down to the roots (Imsande and Touraine, 1994). However, in a split-root experiment with Ricinus neither qualitative nor quantitative changes in the amino acid fraction transported to the roots could be found, even though they contributed differently to N nutrition (Tillard et al., 1998). Although influx studies under the supply of amino acids or metabolic inhibitors repeatedly pointed to glutamine as a major metabolic trigger for the down-regulation of ammonium and nitrate uptake (Rawat et al., 1999; Vidmar et al., 2000), it is possible that this signal became diluted in those experiments. The concentration of a regulatory shoot signal could locally change and then not be detectable in a conventional root analysis. Thus, future analyses might better consider a fractionation of root portions. A direct comparison of ammonium and nitrate influx along the root axis showed that ammonium influx was high and at a similar level within the apical 60 mm of a maize root, except at the very root tip, while nitrate influx increased from apical to basal root zones (Taylor and Bloom, 1998). In the presence of ammonium, nitrate uptake was lowest at the very root tip (Bloom et al., 2003), which might result from repression of nitrate transport at the transcriptional and post-transcriptional level (Orsel et al., 2002).
Alternative phloem-translocated compounds with signalling action, other than amino acids, might be auxin that seems to be involved in the systemic repression of lateral root growth under high N supply (Forde, 2002), and sucrose that might co-ordinate root uptake of both N forms with the C assimilation in the shoot (Lejay et al., 2003).
|
Conclusions |
|---|
(1) Although ammonium is a major N form for plant nutrition and may show a heterogenic distribution in the rooting zone, ammonium-sensory proteins that alter root growth in the presence of ammonium have not yet been uncovered.
(2) AMT-type transporters represent a major path for ammonium influx and mediate NH+4 uniport, while a protein-mediated pathway for the membrane transport of NH3 might occur in parallel, but should play a minor role in ammonium influx.
(3) AMT protein sequences for Arabidopsis, tomato, and rice encompass a different number of homologues and, generally, cluster into two clades, AMT1- and AMT2/MEP-type proteins. In rice, however, a larger number of AMT2-related sequences is found.
(4) AMT genes are regulated at the transcriptional level by N supply, sugars and daytime.
(5) Except for one study, micromolar substrate affinities and a relatively high substrate specificity have been reported for AMT proteins.
(6) Ammonium transport is most likely feedback-regulated at the cellular level by ammonium efflux, compartmentation, and assimilation, but a direct regulatory link between these processes and AMT transporters has not yet been established.
(7) Unlike nitrate, a systemic repression of ammonium influx by the N nutritional status remains to be demonstrated. By contrast, the local N supply to a given root portion, and, in particular, local ammonium and/or glutamine concentrations, exert a strong inhibition of ammonium influx.
|
Acknowledgements |
|---|
We thank Dr Alain Gojon, ENSA Montpellier, France, and Dr Uwe Ludewig, ZMBP Tübingen, Germany, for enjoyable co-operation. We also thank the Deutsche Forschungsgemeinschaft, Bonn, Germany, for financial support from the ‘Schwerpunkt Membrantransport’ (WI1728/4).
|
References |
|---|
Allen S, Smith, JAC. 1986. Ammonium nutrition in Ricinus communis: its effect on plant growth and the chemical composition of the whole plant, xylem sap, and phloem saps. Journal of Experimental Botany 184, 1599–1910.
Becker D, Stahnke R, Fendrick I, Frommer WB, Vanderleyden J, Kaiser WM, Hedrich R. 2002. Expression of the ammonium transporter gene LeAMT1;2 is induced in tomato roots upon association with N2-fixing bacteria. Planta 215, 424–429.[CrossRef][ISI][Medline]
Bloom AJ, Meyerhoff PA, Taylor AR, Rost TL. 2003. Root development and absorption of ammonium and nitrate from the rhizosphere. Journal of Plant Growth Regulation 21, 416–431.
Britto DT, Glass ADM, Kronzucker HJ, Siddiqi MY. 2001a. Cytosolic concentrations and transmembrane fluxes of NH+4/NH3. An evaluation of recent proposals. Plant Physiology 125, 523–526.
Britto DT, Kronzucker HJ. 2001. Constancy of nitrogen turnover kinetics in the plant cell: insights into the integration of subcellular N fluxes. Planta 213, 175–181.[CrossRef][ISI][Medline]
Britto DT, Kronzucker HJ. 2002. NH+4 toxicity in higher plants: a critical review. Journal of Plant Physiology 159, 567–584.[CrossRef][ISI]
Britto DT, Siddiqi MY, Glass ADM, Kronzucker HJ. 2001b. Futile transmembrane NH+4 cycling: a cellular hypothesis to explain ammonium toxicity in plants. Proceedings of the National Academy of Sciences, USA 98, 4255–4258.
Brüggemann L, Dietrich P, Becker D, Dreyer I, Palme K, Hedrich R. 1999. Channel-mediated high-affinity K+ uptake into guard cells from Arabidopsis. Proceedings of the National Academy of Sciences, USA 96, 3298–3302.
Causin HF, Barneix AJ. 1993. Regulation of NH+4 uptake in wheat plants: effects of root ammonium concentration and amino acids. Plant and Soil 151, 211–219.[CrossRef]
Colmer TD, Bloom AJ. 1998. A comparison of net NH+4 and NO–3 fluxes along roots of rice and maize. Plant, Cell and Environment 21, 240–246.
Coruzzi G, Bush DR. 2001. Nitrogen and carbon nutrient and metabolite signaling in plants. Plant Physiology 125, 61–64.
Coutts G, Thomas G, Blakey D, Merrick M. 2002. Membrane sequestration of the signal transduction protein GlnK by the ammonium transporter AmtB. EMBO Journal 21, 536–545.[CrossRef][ISI][Medline]
Demidchik V, Tester M. 2002. Sodium fluxes through nonselective cation channels in the plasma membrane of protoplasts from Arabidopsis roots. Plant Physiology 128, 379–387.
Dennison KL, Robertson WR, Lewis BD, Hirsch RE, Sussman MR, Spalding EP. 2001. Functions of AKT1 and AKT2 potassium channels determined by studies of single and double mutants of Arabidopsis. Plant Physiology 127, 1012–1019.
Drew MC. 1975. Comparison of the effects of a localized supply of phosphate, nitrate, ammonium, and potassium on the seminal root system and the shoot in barley. New Phytologist 75, 479–490.[CrossRef][ISI]
Dubois E, Grenson M. 1979. Methylamine/ammonia uptake systems in Saccharomyces cerevisiae: multiplicity and regulation. Molecular and General Genetics 175, 67–76.
Forde BG. 2002. The role of long-distance signalling in plant responses to nitrate and other nutrients. Journal of Experimental Botany 53, 39–43.
Fu HH, Luan S. 1998. A dual-affinity K+ transporter from Arabidopsis. The Plant Cell 10, 63–73.
Gansel X, Munos S, Tillard P, Gojon A. 2001. Differential regulation of the NO–3 and NH+4 transporter genes AtNrt2.1 and AtAmt1.1 in Arabidopsis: relation with long-distance and local controls by N status of the plant. The Plant Journal 26, 143–155.[CrossRef][ISI][Medline]
Gassmann W, Schroeder JI. 1994. Inward-rectifying K+ channels in root hairs of wheat. Plant Physiology 105, 1399–1408.[Abstract]
Gazzarrini S, Lejay L, Gojon A, Ninnemann O, Frommer WB, von Wirén N. 1999. Three functional transporters for constitutive, diurnally regulated, and starvation-induced uptake of ammonium into Arabidopsis roots. The Plant Cell 11, 937–947.
Glass ADM, Britto DT, Kaiser BN, et al. 2002. The regulation of nitrate and ammonium transport systems in plants. Journal of Experimental Botany 53, 855–864.
Hanstein S, Felle HH. 1999. The influence of atmospheric NH3 on the apoplastic pH of green leaves: a non-invasive approach with pH-sensitive microelectrodes. New Phytologist 143, 333–338.[CrossRef]
Husted S, Mattsson M, Mollers C, Wallbraun M, Schjoerring JK. 2002. Photorespiratory NH+4 production in leaves of wild-type and glutamine synthetase 2 antisense oilseed rape. Plant Physiology 130, 989–998.
Imsande J, Touraine B. 1994. N demand and the regulation of nitrate uptake. Plant Physiology 105, 3–7.[ISI][Medline]
Ishiyama K, Kojima S, Takahashi H, Hayakawa T, Yamaya T. 2003. Cell type distinct accumulations of mRNA and protein for NADH-dependent glutamate synthase in rice roots in response to the supply of NH+4. Plant Physiology and Biochemistry 41, 643–647.[CrossRef]
Kaiser BN, Rawat SR, Siddiqi MY, Masle J, Glass ADM. 2002. Functional analysis of an Arabidopsis T-DNA ‘Knockout’ of the high-affinity NH+4 transporter AtAMT1;1. Plant Physiology 130, 1263–1275.
Kiyomiya S, Nakanishi H, Uchida H, et al. 2001. Real time visualization of N13-translocation in rice under different environmental conditions using positron emitting tracer imaging system. Plant Physiology 125, 1743–1753.
Kosola KR, Bloom AJ. 1994. Methylammonium as a transport analog for ammonium in tomato (Lycopersicon esculentum). Plant Physiology 105, 435–442.[Abstract]
Kronzucker HJ, Schjoerring JK, Erner Y, Kirk GJD, Siddiqi MY, Glass ADM. 1998. Dynamic interactions between root NH+4 influx and long-distance N translocation in rice: Insights into feedback processes. Plant and Cell Physiology 39, 1287–1293.
Kumar HJ, Silim SN, Okamoto M, Siddiqi MY, Glass ADM. 2003. Differential expression of three members of the AMT1 gene family encoding putative high-affinity NH+4 transporters in roots of Oryza sativa sub