Futile transmembrane NH₄⁺ cycling: A cellular hypothesis to explain ammonium toxicity in plants

Abstract

Most higher plants develop severe toxicity symptoms when grown on ammonium (NH) as the sole nitrogen source. Recently, NH toxicity has been implicated as a cause of forest decline and even species extinction. Although mechanisms underlying NH toxicity have been extensively sought, the primary events conferring it at the cellular level are not understood. Using a high-precision positron tracing technique, we here present a cell-physiological characterization of NH acquisition in two major cereals, barley (Hordeum vulgare), known to be susceptible to toxicity, and rice (Oryza sativa), known for its exceptional tolerance to even high levels of NH. We show that, at high external NH concentration ([NH]_o), barley root cells experience a breakdown in the regulation of NH influx, leading to the accumulation of excessive amounts of NH in the cytosol. Measurements of NH efflux, combined with a thermodynamic analysis of the transmembrane electrochemical potential for NH, reveal that, at elevated [NH]_o, barley cells engage a high-capacity NH-efflux system that supports outward NH fluxes against a sizable gradient. Ammonium efflux is shown to constitute as much as 80% of primary influx, resulting in a never-before-documented futile cycling of nitrogen across the plasma membrane of root cells. This futile cycling carries a high energetic cost (we record a 40% increase in root respiration) that is independent of N metabolism and is accompanied by a decline in growth. In rice, by contrast, a cellular defense strategy has evolved that is characterized by an energetically neutral, near-Nernstian, equilibration of NH at high [NH]_o. Thus our study has characterized the primary events in NH nutrition at the cellular level that may constitute the fundamental cause of NH toxicity in plants.

Plants can extract and use various forms of nitrogen (N) from soils, most importantly the inorganic ions ammonium (NH) and nitrate (NO). While one might expect NH to be preferred by plants, as its assimilation requires less energy than that of NO (1), only a few species perform well when NH is the only, or predominant, source of N (2). By contrast, most species develop toxicity symptoms when grown on moderate to high levels of NH (1–6), whereas normal growth in these species is seen on NO. Ammonium toxicity is especially problematic in areas with intensive agriculture and cultivation of livestock, where high levels of NH₃ emission, and subsequent NH deposition, are observed (6, 7). It is estimated that N deposition from atmospheric NH₃ can reach levels as high as 50 kg⋅ha⁻¹⋅yr⁻¹ in some parts of Europe, in extreme cases constituting 50–80% of total N deposited from all possible sources (6). As a result, N saturation in many natural ecosystems is exceeded by as much as 10-fold, and damage to forest and agricultural crops alike has been attributed directly to this phenomenon (8, 9). This is a problem of serious concern in both Europe and North America (6–10).

Several hypotheses have been advanced to explain why NH is toxic to plants, but none is considered satisfactory (3, 5). As NH uptake mechanisms are coupled to H⁺ extrusion into the rooting medium and H⁺ release is also associated with NH incorporation into protein (6), it has been repeatedly suggested that root medium acidification (11) and/or intracellular pH disturbance (3, 12) may explain the observed symptoms. However, in many cases, toxicity is equally observed in pH-buffered media (13), and a recent study on pea (Pisum sativum), known to be ammonium sensitive, has discounted the occurrence of ammonium-induced cytosolic pH disturbance (12). Others have suggested that carbohydrate limitation may contribute to the toxicity syndrome, based on the finding that NH per se is not translocated to the shoot in most plants (14), and, thus, all C skeletons for N assimilation must be provided in roots, causing local C deprivation (15). In some cases, external provision of α-ketoglutarate, a key carbon source for N assimilation, alleviated toxicity symptoms (4), but in other cases it failed to enhance NH metabolism (16), suggesting that other factors may limit NH assimilation. The hypothesis that NH toxicity results from the uncoupling of photophosphorylation in chloroplasts (6) has long been shown to be incorrect, as even very large NH concentrations do not affect this process in intact chloroplasts (17). A diminution of essential cations such as K⁺, Mg²⁺, and Ca²⁺ in tissues of plants grown under prolonged exposure to NH has been widely documented (18–20), and while this may be a contributor to the toxicity syndrome, it is clearly a longer-term effect, resulting from nutrient imbalance.

What these hypotheses have in common is that they are all contingent upon the permeation of NH into the cell. Therefore, the study of transmembrane NH fluxes is essential to the establishment of a proper context for the discussion of the issue of toxicity. Surprisingly, however, no study of the primary events of NH acquisition, including the modes of entry, metabolism, and sequestration, has been undertaken in this context. Here we have used positron-emission tracing with the short-lived nitrogen isotope ¹³N to characterize these primary events at the cellular level, in the model grass species rice (Oryza sativa) and barley (Hordeum vulgare).

Methods

Plant Culture and Radiotracer Experiments.

Seedlings of barley (cv. CM-72) and of two major tropical-wetland varieties of rice (cv. IR-72 and cv. M-202) were cultivated hydroponically for 7 d and 21 d, respectively, in controlled-environment chambers. Hydroponic tanks contained aerated and N-free modified Johnson's solution (2, 14, 16). NH was supplied as (NH₄)₂SO₄ at steady-state concentrations of 0.1 mM or 10 mM during growth and for radiolabeling experiments, which were conducted in the growth chambers. The short-lived radiotracer ¹³N (half-life = 9.98 min) was used to trace NH fluxes and to estimate cytosolic concentrations of NH by compartmental analysis, as described previously (2, 16, 21). In brief, roots of intact seedlings were exposed to growth solution spiked with ¹³NH for 1 h, to maximize the specific activity of tracer in the cytosolic compartment of root cells (21). Seedlings were then transferred to efflux funnels (16, 21), and tracer from the roots was eluted successively with 20-ml aliquots of nonlabeled solution for various washout periods. With t = 0 as the time of transfer from loading to washing solution and t_final = 24 min for the final elution, the time periods for the 25 successive washes were: 10 s (three times), 15 s (six times), 30 s (four times), 1 min (four times), and 2 min (eight times). γ-Ray emissions in the eluates were then measured (see Fig. 1). After the final elution, roots were excised from shoots, spin-dried for 30 s, and weighed, and the γ-ray emissions of roots and shoots were counted. Influx, efflux, net flux, and cytosolic concentrations of NH were estimated from analysis of the wash-out kinetics as described in detail elsewhere (2, 16, 21). All fluxes are expressed in μmol per g (fresh weight) per h. Symbols, and basic calculation methods, for fluxes are as follows:

Figure 1.

Semilogarithmic plots of ¹³NH efflux from the cytosolic compartments of barley and rice roots. Plants were prelabeled under steady-state conditions, with NH provided externally at 10 mM. Plots have been corrected for specific activities of ¹³N tracer (2, 21), allowing a direct comparison of initial efflux rates by inspection of the y-intercepts of the regression lines.

Φ_co = efflux from the cytosol, obtained from the rate of ¹³N release from the cytosol at time 0.

Φ_net = net flux, obtained directly from the accumulation of ¹³N in the plants at the end of the elution period.

Φ_oc = unidirectional influx, calculated from Φ_net + Φ_co.

Φ_xylem = xylem flux of ¹³N, obtained directly from accumulation in the shoot, at the end of the elution period.

Φ_met/vac = combined flux to assimilation and vacuole, resulting from Φ_net − Φ_xylem.

Cytosolic NH concentrations were calculated from the quotient of the rate of ¹³NH release integrated over 5 times the half-life of cytosolic NH exchange, and the ratio of efflux to all fluxes removing ¹³NH from the cytosol, and were based on the assumption that the cytosol occupies 5% of cell volume (2, 16, 21). Experiments were repeated four to five times. Standard errors (SE) for fluxes were less than 20% of the means.

Electrophysiological Measurements.

Membrane potential differences were measured as described elsewhere (22). Intact barley or rice roots were held between steel pins and pieces of silicone tubing in a Plexiglas chamber affixed to a light microscope, and they were impaled with glass microelectrodes until a stable electrical potential difference between root and external solution was maintained for at least 2 min. Sixty-seven measurements in total were made on different plants for the 0.1 mM NH condition in barley, and 20 for the 10 mM NH condition. SE was <3% of the means. For rice, 17 measurements were undertaken at 0.1 mM external NH concentration ([NH]_o), and 26 at 10 mM [NH]_o (SE < 5% of the means).

Respiration Measurements.

O₂ depletion was followed potentiometrically in intact roots of barley and rice with a Hansatech cuvette/electrode system attached to a chart recorder. Cuvettes were filled with air-saturated growth solutions. Calculations were based on slopes of steady O₂ depletion lines generated over 3–10 min, after which plants were weighed and discarded. The O₂ content of the cuvette was not reduced by more than 25% of the initial value. Measurements were repeated 20–25 times.

Relative Growth Rates.

Individual barley plants were monitored over days 1, 3, and 7 after germination. Surface water was removed from roots before weighing, after which the plants were immediately returned to solution. Eight replicates were used for each of the two steady-state NH-growth concentrations.

Results and Discussion

In the present study, we assayed NH transport properties in the root cells of two major cereal species known to differ markedly in their abilities to tolerate NH as the sole N source. ¹³NH efflux from roots of intact plants of the two species, prelabeled with this tracer, showed substantial differences in the rates of NH extrusion from root cells to the external medium (Fig. 1). Interestingly, barley (Hordeum vulgare), which suffers from NH toxicity (23), excretes NH at a significantly higher rate than rice (Oryza sativa), a species whose tolerance to NH is considered exceptional (24). Fig. 2, depicting bidirectional chemical fluxes of NH across the plasma membranes of these two species, documents in both cases a large increase in total flux (equivalent to unidirectional influx) as the steady-state external NH concentration is stepped up from a moderate level of 0.1 mM to a level of 10 mM, which represents the high end of NH concentrations found in agricultural soils (1–11), and which is potentially toxic (24). Importantly, however, the increase in barley is much higher (12-fold) than in rice (7-fold). In fact, the magnitudes of the fluxes in barley are the highest ever reported for NH in a plant root system, and strongly suggest the operation of channel-based NH-transport systems, whose existence has been demonstrated by several recent studies (25, 26). As the steady-state net fluxes are almost identical in the two species, a substantially larger fraction of incoming nitrogen is subsequently excreted in the ammonium-sensitive species. This pronounced excretion of N is inextricably linked to an apparent inability of barley to exclude NH at the more primary, intake, step. Our preliminary analyses on other species known to suffer NH toxicity, including wheat, tomato, and the tree species Douglas-fir and trembling aspen (27), suggest that this phenomenon is by no means isolated, but may well occur universally among susceptible higher plants.

Figure 2.

Comparison of steady-state bidirectional plasma-membrane NH fluxes in barley and rice roots at 0.1 mM and 10 mM [NH]_o. Column height represents total influx from the external medium (Φ_oc), while the filled areas depict the portion of influx returned to the environment by efflux transport (Φ_co). The net flux (Φ_net) is the difference between these two fluxes. Vertical bars indicate standard errors of influx means.

Knowledge of the kinetic constants of NH exchange for the cytosolic component in the experimental root system, and of the tracer content in this pool at the onset of the elution protocol, allowed us to compare the concentrations of NH on either side of the plasma membrane (2), and hence to investigate the energetics of fluxes across this membrane. Our analysis shows that, at 10 mM [NH]_o, the cytosolic concentrations of NH in both barley and rice are of a level higher than ever reported in plant cells (Table 1), and indeed are irreconcilable with the widely held dogma that no appreciable quantities of free NH can exist in the cytosol (1, 5, 6). Nevertheless, a thermodynamic analysis, using measurements of electrical potential (ΔΨ) across the plasma membrane, reveals that, in barley, the cytosolic levels of NH at 10 mM [NH]_o are substantially lower than predicted by the Nernst equation, which applies when NH is passively distributed across a highly permeable membrane (Table 1). Rice, by contrast, maintains a lower internal concentration than barley under this condition, but because of the relatively depolarized state of its plasma membrane under high NH provision (22), the transmembrane distribution ratio of NH very closely approximates that predicted by Nernstian analysis (Table 1). This lowering of membrane polarization with increasing NH provision, which is not followed by a restoration of that polarization in the steady state, is confirmed by previously published work on rice (ref. 22; see Table 1) and the aquatic NH-specialist plant Lemna gibba (28). Table 1 shows that, in rice, a depolarized membrane potential is already achieved at 1 mM [NH]_o, a more modest NH concentration. The reduction of ΔΨ with increasing [NH]_o apparent in rice has the important biophysical consequence of lowering the ceiling for NH accumulation in the cytosol, thus eliminating the gradient against which efflux transporters must work to remove excess cytosolic NH. The failure of barley to down-regulate ΔΨ, on the other hand, sustains a gradient for very large inward NH fluxes, which we in fact observe (Fig. 2). Homeostatic restoration of ΔΨ to preset values has been previously reported for barley exposed to various concentrations of NO (29), the second major N source used by plants. Under NH nutrition, an undiminished ΔΨ in barley yields a high potential for NH accumulation, and a significantly larger demand on the efflux process (Table 1). Our data show that, whereas the net flux of NH is nearly identical in both species at 10 mM, the efflux process in barley mediates a 2- to 3-fold higher removal of NH from the cytosol, effectively reducing a Nernstian NH concentration of 1.32 M by the equivalent of 962 mM (Table 1).

Table 1.

Thermodynamic analysis of cytosolic NH pool sizes in barley and rice

Plant	[NH]o, mM	Influx, μmol⋅g−1⋅h−1	Efflux, μmol⋅g−1⋅h−1	Flux ratio	ΔΨ, mV	[NH]c, mM
						Predicted	Measured
Barley	0.1	5.46	1.53	0.28	−121	12	28
	10	65.1	49.4	0.76	−123	1,320	358
Rice	0.1	4.85	0.67	0.14	−132	15.6	33.4
		5.97 (ref. 34)	1.17 (ref. 34)	0.19	−122 (ref. 22)	10.7	20.6 (ref. 34)
	1	10.51 (ref. 34)	3.09 (ref. 34)	0.29	−88 (ref. 22)	28.78	38.1 (ref. 34)
	10	35.0	18.6	0.53	−87.9	289	232
					−82 (ref. 22)	233	ND

The Nernst equation {Ψ_N = RT/(nF) ln([NH]_o/[NH]_c), where Ψ_N = the Nernst potential; R = 8.3144 J⋅K⁻¹⋅mol⁻¹; F = 96,485 C⋅mol⁻¹; n = 1; and T = 293.15 K (for barley) or 303.15 K (for rice)} was used to predict equilibrium cytosolic concentrations ([NH]_c) on the basis of external concentration, membrane potential (ΔΨ), and an ambient temperature of 20°C for barley and 30°C for rice. Cytosolic concentrations were measured by using compartmental analysis (see text). Additional membrane potential readings and estimates of cytosolic NH concentrations for rice (cv. M-202) are included on the basis of analysis of previously published work (22, 34). ND, not determined. 

Such a process must carry a substantial energetic burden, and this prediction was consistent with measurements of oxygen consumption in roots of intact barley plants under the steady-state conditions of NH supply corresponding to those of flux measurements. As indicated in Fig. 3, upon transition from low steady-state NH provision to high, respiration in barley increased by a startling 41%, whereas no significant difference was found in rice between growth conditions (not shown). To isolate, in barley, the flux processes at the plasma membrane level from other potentially energy-requiring pathways of NH processing, we applied methionine sulfoximine (MSX), a compound known to completely block NH assimilation in the species under investigation (14, 16, 30). Importantly, MSX did not diminish respiratory activity associated with elevated external NH (Fig. 3), emphasizing that the respiratory stimulation was attributable to processes upstream of NH metabolism—i.e., residing at the level of membrane transport—and independent of activities of glutamine synthetase and other processes related to N metabolism. Other work on barley has shown that the proportion of total respiration assigned to processes associated with NH uptake and assimilation constitutes as much as 14% of the plant's total respiratory expenditure at 0.1 mM external NH provision (31). Most of this expenditure appears to be due to plasma-membrane transport activity, hence we conclude that the 41% increase in total root respiration observed in the present study for barley at 10 mM NH can be attributed directly to the approximately 30-fold increase in NH efflux from the cytosol under this condition. Work with the NH analogue methylammonium shows that a respiratory increase is also observed in barley (32), as well as other species (33), on addition of this nonmetabolized compound. In the absence of metabolic sinks, it becomes clear that membrane processes per se can impose a substantial energetic burden on the plant root system. Indeed, in barley the increase in respiration at 10 mM external NH is not correlated with an increase in growth-related energy demands; on the contrary, it accompanies a depression in relative growth rate at the whole plant level (Fig. 3 Inset) not seen in ammonium-tolerant species such as rice (2, 14, 21, 22, 34, 35). It is intriguing that in rice, where no respiratory increase was found with increasing NH provision (data not shown), a thermodynamic equilibrium is achieved across the plasma membrane (Table 1), suggesting that passive inwardly and outwardly directed NH channel activities may mediate NH distribution across this membrane without energetic cost. These findings underscore the cellular adaptations rice has evolved to use NH as an N source (34, 35), which allow it to thrive in flooded soil environments where hypoxic to anoxic conditions typically render NH the only major N source available to plant growth (35).

Figure 3.

Respiration rates of intact barley roots at 0.1 mM and 10 mM [NH]_o. In one experiment, 1 mM methionine sulfoximine (MSX) was applied to block NH metabolism (see text). Relative growth rates under the two conditions are shown in the Inset. Rice experienced no significant difference in respiration under the two NH regimes (see text).

We propose that the inability of barley, and that of other species susceptible to ammonium toxicity, to exclude NH by regulation of plasma-membrane influx systems constitutes a fundamental breakdown in plant cell function and must precede any intracellular toxicity-associated events such as cation displacement or carbohydrate depletion (see above). The operation of the energy-intensive NH extrusion mechanism we describe appears to be central to the ammonium toxicity syndrome, and is similar in principle to mechanisms evolved by bacteria and carcinomas to actively excrete cytotoxins such as antibiotics and chemotherapeutic agents (36, 37). The ecological significance of this syndrome is substantial, not only because soil nitrogen profiles profoundly determine the spatial and temporal dynamics of ecosystems (2, 38), but especially in the light of recent interventions by humans in the functioning of the global nitrogen cycle (5–7, 39).

Abbreviation

[NH]_o

external NH concentration