Wheat leaves emit nitrous oxide during nitrate assimilation

Abstract

Nitrous oxide (N₂O) is a key atmospheric greenhouse gas that contributes to global climatic change through radiative warming and depletion of stratospheric ozone. In this report, N₂O flux was monitored simultaneously with photosynthetic CO₂ and O₂ exchanges from intact canopies of 12 wheat seedlings. The rates of N₂O-N emitted ranged from <2 pmol⋅m⁻²⋅s⁻¹ when NH was the N source, to 25.6 ± 1.7 pmol⋅m⁻²⋅s⁻¹ (mean ± SE, n = 13) when the N source was shifted to NO. Such fluxes are among the smallest reported for any trace gas emitted by a higher plant. Leaf N₂O emissions were correlated with leaf nitrate assimilation activity, as measured by using the assimilation quotient, the ratio of CO₂ assimilated to O₂ evolved. ¹⁵N isotopic signatures on N₂O emitted from leaves supported direct N₂O production by plant NO assimilation and not N₂O produced by microorganisms on root surfaces and emitted in the transpiration stream. In vitro production of N₂O by both intact chloroplasts and nitrite reductase, but not by nitrate reductase, indicated that N₂O produced by leaves occurred during photoassimilation of NO in the chloroplast. Given the large quantities of NO assimilated by plants in the terrestrial biosphere, these observations suggest that formation of N₂O during NO photoassimilation could be an important global biogenic N₂O source.

Plants play a critical role in regulating the chemical and physical state of the atmosphere through the exchange of biogenic greenhouse gases. Most notable are plant–atmosphere exchanges of CO₂, O₂, and H₂O, but leaves also emit a variety of carbon- and nitrogen-based trace gases involved in climate alteration processes (1). One such trace gas is nitrous oxide (N₂O). Plants—either aerenchymous (2) or nonaerenchymous (3, 4)—can serve as conduits for N₂O between the soil and atmosphere. They transpire significant quantities of N₂O when its concentration in the soil solution greatly exceeds the solution equilibrium concentration with ambient N₂O, currently at ≈315 nmol⋅mol⁻¹ (5).

The global N₂O budget is beset by uncertainty, and sources of N₂O have historically fallen short of the primary sink, photolysis in the upper atmosphere (6–9). The primary biogenic N₂O sources are from soils (70%) and involve the microbial nitrogen transformations brought about by nitrification and denitrification (6). Although nitrification and denitrification are the major N₂O sources, several microbial organisms that do not nitrify or denitrify can also produce N₂O during NO assimilation (10, 11). These observations have led to the general hypothesis that any enzymatic nitrogen transformation through the +2 to +1 oxidation state may generate N₂O (12). One such transformation in higher plants is NO assimilation in chloroplasts. Nitrite assimilation in chloroplasts can generate intermediates capable of reacting to produce N₂O, including NO (as HNO₂) with hydroxylamine (13) or reaction of NO released during NO reduction (14, 15) with ascorbate (16). Nonetheless, early attempts to observe N₂O production by higher plant tissues were not successful (10) and were probably limited by lack of an analytical method capable of detecting plant N₂O emission at the exceptionally slow rates reported here. We developed an analytical approach by using cryogenic trapping (17) and gas chromatography coupled to high-precision isotope ratio mass spectrometry (18). This approach resolved leaf N₂O emissions at more than six orders of magnitude lower than photosynthetic gas exchanges of CO₂ and O₂ (Table 1), placing such fluxes among the smallest ever reported for any trace gas emitted by a higher plant (1).

Table 1.

Shoot and rhizosphere gas exchange rates for 14-day-old wheat seedlings (T. aestivum L. cv. Veery 10)

N source	Photosynthesis			N2O-N emissions
	CO2, μmol⋅m−2⋅s−1	O2, μmol⋅m−2⋅s−1	AQ, CO2/O2	Shoot, pmol⋅m−2⋅s−1	Rhizosphere, pmol⋅g−1⋅s−1
50 μM (15NH4)2SO4	14.31  ± 1.32	12.06  ± 0.77	1.21  ± 0.06	ND	40.4  ± 2.1
100 μM K15NO3	15.32  ± 1.48	13.63  ± 0.84	1.13  ± 0.05	25.59  ± 1.68	304.1  ± 93.8

Shown for shoots are net photosynthetic CO₂ assimilation rates, net photosynthetic O₂ evolved, AQ (the ratio of CO₂ assimilated to O₂ evolved), and shoot N₂O emissions (mean ± SE, n = 11). Shown for the rhizosphere are N₂O production rates (n = 7). 

Identifying and quantifying plant N₂O exchange is important. Atmospheric N₂O concentration is increasing at a rate of about 0.27% per year (19), and each mol of N₂O has ≈290 times the radiative forcing potential of CO₂ (20). Consequently, N₂O will account for as much as 7% of projected atmospheric warming (21). In addition to its greenhouse gas properties, photolytic reaction with excited oxygen [O(¹D)] in the upper atmosphere produces nitric oxide (NO), and NO, in turn, consumes stratospheric ozone (22). Biogenic and anthropogenic sources of N₂O are poorly constrained (23) and often do not account for the quantity of N₂O known to undergo photolysis in the upper atmosphere (6). Extreme heterogeneity of soil N₂O emissions largely contributes to such uncertainty, but unidentified hydrologic or biogenic sources may also play a role (7, 8). Kroeze et al. (24) have argued for closure of the global N₂O budget, but the theoretical uncertainty range for estimates of N₂O from agricultural soils is extreme [0.6–14.8 Tg N₂O-N y⁻¹ (6)], and new evidence suggests that N₂O emissions from agricultural soils may be in the lower range of recent Intergovernmental Panel on Climate Change (IPCC) estimates (25). To moderate the atmospheric increases of N₂O and to better understand the role of the biosphere in its production, it is critical to identify all major N₂O sources and exchange pathways, including any potential contributions by plants.

Methods

Wheat seeds (Triticum aestivum L. cv. Veery 10) were surface sterilized, germinated on rolled moist germination paper, and transferred to opaque hydroponics tanks containing dilute nutrient solutions (26). The hydroponics systems were kept in a controlled environment chamber (PGV36, Conviron, Winnipeg, MB, Canada) with a 25°C, 16-h, 600 μmol⋅m⁻²⋅s⁻¹ photosynthetic photon flux density day and a 20°C, 8-h night. After 10 days, when the plants had three true leaves, 12 individuals were transferred into a gas exchange chamber where their shoots produced a canopy with a leaf surface area of about 0.02 m². Roots were sealed into 12 individual gas-tight cuvettes that were connected to the lower surface of a platform over which the canopy chamber was sealed (27).

Net canopy CO₂ and O₂ exchanges were measured under steady-state conditions. A differential infrared gas analyzer (VIA-500R, Horiba, Sunnyvale, CA) monitored net CO₂ exchange. From a second parallel gas stream, a custom oxygen analyzer that resolves a 2 μmol⋅mol⁻¹ O₂ partial pressure difference on a background of 209,460 μmol⋅mol⁻¹ monitored net O₂ exchange (28). Before passing through the O₂ electrodes, gases that can interfere with O₂ measurements, like water vapor and plant secondary carbon compounds, were cryogenically condensed to constant trace levels by using a liquid argon trap. The CO₂ concentration during the gas exchange measurements was sustained at 330 μmol⋅mol⁻¹ and the N₂O concentration at 280 nmol⋅mol⁻¹. A third parallel gas stream exiting the plant chamber flowed at 1.0–1.5 dm³ s⁻¹ through an ascarite filter to remove CO₂, a drierite filter to remove H₂O, and then a second cryogenic trap cooled with liquid argon to condense N₂O (17). This trap concentrated up to 10 μmol⋅mol⁻¹ N₂O in 15 min when the N₂O concentration in the gas stream was at 280 nmol⋅mol⁻¹. Gas trapped in the condenser was injected into a mass spectrometer tuned to determine the ratio of ion currents m/z at 45/44 and 46/44 (18). The rate of N₂O exchange by shoots (mol⋅m⁻² s⁻¹) was calculated as F_N2O = (J_l⋅d_l⋅C_l/A) (29). J_l is the flow rate of air through the plant chamber (mol⋅s⁻¹). d_l is the mol fraction of N₂O proportional to canopy N₂O emission, calculated according to the isotopic enrichments in masses ⁴⁵N₂O and ⁴⁶N₂O (30). C_l is the concentration of N₂O in the gas stream exiting the chamber. A is leaf area (m²). The theoretical detection limit for leaf N₂O emission by the mass spectrometer, on the basis of the observed variation in the mass ratios reported for a 1.5 μmol⋅mol⁻¹ N₂O standard, was ≈2 pmol⋅m⁻²⋅s⁻¹.

During the first 24 h of our experiments, roots received a nitrogen source containing 50 μM (¹⁵NH₄)₂SO₄ (99.6 atom % ¹⁵N), and shoot gas fluxes of CO₂, O₂ and N₂O were assessed during the final 6 h. This pretreatment allowed us to purge the xylem stream of NO and establish a baseline value for assimilation quotient (AQ, CO₂/O₂) when little NO was undergoing assimilation in leaves. The nitrogen source was then shifted to 100 μM K¹⁵NO₃ (99.6 atom % ¹⁵N) for 24 h, and shoot gas fluxes were again assessed during the final 6 h.

For rhizosphere N₂O production, the nutrient solution was delivered to roots by using a gas tight, continuous flow system (31). Fluxes of N₂O from the root cuvettes represent the production of N₂O by roots and any microbial organisms on the root surface capable of generating N₂O (32). Before passing through the root cuvettes, the solution concentrations of O₂ and N₂O were brought to their respective saturation concentrations. Dissolved oxygen concentration in the nutrient solution passing through the root cuvettes declined by only about 20% and the solutions were well stirred, so denitrification in the rhizosphere was minimized. During the experiments, two 5-ml samples of nutrient solution were collected into 15-ml septum bottles. One sample was collected before the nutrient solution entered the root cuvette and another after it exited the root cuvette. The head space gases from the septum bottles were injected into the mass spectrometer after equilibration at 22.5°C to determine the concentration of N₂O and the ratio of ion currents m/z at 45/44 and 46/44. The rate of N₂O production by the rhizosphere (mol⋅g⁻¹⋅s⁻¹) was calculated as F_N2O = (J_r⋅d_r⋅C_r/W). J_r is the flow rate of nutrient solution through the root cuvette (liter⋅s⁻¹). d_r is the mol fraction of N₂O in the nutrient solution proportional to rhizosphere N₂O production, and calculated according to the isotopic enrichments in masses ⁴⁵N₂O and ⁴⁶N₂O (30). C_r is the total concentration of N₂O (mol⋅liter⁻¹) in the nutrient solution (33). W is root dry mass.

For chloroplast assays, ≈40 g of fresh leaves from 2-week-old hydroponics-grown wheat plants was blended in a buffer solution containing 0.05 M K-Hepes (pH 7.3), 0.33 M Sorbitol, 1 mM MgCl₂, 1 mM MnCl₂, 2 mM Na₂EDTA, and 0.1% BSA. The extract was centrifuged at 3,000 × g for 5 min, resuspended in a 50/50 Percol gradient, and centrifuged at 7,000 × g for 10 min. Intact chloroplasts were then collected and washed with 0.05 M K-Tricine (pH 8.0) and 0.33 M Sorbitol. The washed chloroplasts were introduced into 5 ml of an incubation medium consisting of 0.3 mM K¹⁵NO₂ (99.6 atom % ¹⁵N) in 0.05 M K-Tricine (pH 8.0), 0.33 M sorbitol, and 0.3 mM NaHCO₃. The chloroplasts were incubated for 40 min in 15-ml septum bottles at 25°C and a light intensity of 600 μmol⋅m⁻²⋅s⁻¹ photosynthetic photon flux density. At the end of the incubation period, the head-space gases were collected and injected into the mass spectrometer. The concentration of N₂O and ratio of ion currents m/z at 45/44 and 46/44 were used to calculate the quantity of N₂O produced, as previously described for the dynamic flow measurements.

For nitrate and nitrite reductase assays (34), 2 g of fresh leaves from 2-week-old hydroponics-grown wheat plants was ground in a mortar and pestle with cold-purified N-free sand in 8 ml of a chilled buffer solution. The buffer solution consisted of 0.05 M Tris (pH 8.5), 1 mM EDTA, 1 μM NaMoO₄, 10 μM FAD, 1 mM DTT, 10 μM leupeptin, and 1 μg ml⁻¹ pepstatin. The extract was centrifuged at 30,000 × g for 20 min. Then 0.05 ml of the supernatant was added to 0.95 ml of an assay solution containing 62.5 mM potassium phosphate buffer (pH 7.5), 0.7 mM K¹⁵NO₂ (99.6 atom % ¹⁵N), and 0.052 g ml⁻¹ of methyl viologen in a 15-ml septum bottle. The reaction was initiated by injecting 0.2 ml of a solution containing 8.3 mg ml⁻¹ Na₂S₂O₄, and incubated at 30°C for 15 min. For nitrate reductase, 0.05 ml of the supernatant was added to 0.95 ml of assay buffer containing 1.4 mM K¹⁵NO₃ (99.6 atom % ¹⁵N). The reaction was initiated by adding 0.2 ml of a solution containing 2 mg ml⁻¹ NADH and incubated at 30°C for 15 min. The head-space gases from the nitrite and nitrate reductase assays were collected and analyzed on the mass spectrometer as previously described.

Results and Discussion

This is the first study, to our knowledge, to report quantitative leaf N₂O emissions under normal physiological conditions for an intact plant by using steady-state gas exchange methods. Earlier investigations have examined detached leaves or used static chamber methods, where CO₂, H₂O, and temperature change rapidly (3, 4, 10), and nitrogen source (NH and NO) cannot be controlled. These preliminary studies were valuable in that they clearly demonstrated N₂O can move from soil to atmosphere via the plant transpiration stream, but they did not resolve the question of N₂O production by plant nitrogen metabolism.

In our investigation, leaves did not emit N₂O at a detectable rate while metabolizing ¹⁵NH. The AQ during ¹⁵NH exposure averaged 1.21 ± 0.05 units (Table 1). When the N source was shifted to ¹⁵NO, canopy leaves emitted N₂O-N at an average rate of 25.6 ± 1.7 pmol⋅m⁻²⋅s⁻¹, and the AQ declined to 1.13 ± 0.06 units (Table 1). Both observations are important. The change in AQ (ΔAQ) gives a nondestructive measure of NO assimilation under steady-state conditions. Net O₂ exchange provides a measure of photosynthetic electron transport, and either CO₂ or NO reduction can be coupled to such electron transfer (35). As NO photoassimilation increases, the AQ declines because reductant produced by photosynthetic electron transport increases to support NO and NO reduction in addition to CO₂ fixation (28, 36). Thus, not only did shoot N₂O emission become detectable when the N source was shifted to ¹⁵NO, but the strong correlation with ΔAQ (Fig. 1A) suggested that leaf N₂O flux was driven by leaf NO assimilation.

Figure 1.

The relationship between N₂O emission from wheat leaves (pmol N₂O-N m⁻²⋅s⁻¹) and (A) the change in assimilatory quotient (ΔAQ) when the nitrogen source was shifted from ¹⁵NH to ¹⁵NO, or (B) N₂O production in the rhizosphere (pmol N₂O-N g⁻¹⋅s⁻¹). Shown are the regression lines and the R² statistic. The rates of N₂O emission were normalized by using the log₁₀ transformation.

On the other hand, rhizosphere N₂O production also increased more than 7-fold (Table 1) with the change in N source from ¹⁵NH to ¹⁵NO (Table 1), raising N₂O concentration in the nutrient solution around roots to an average of 89.6 nmol⋅liter⁻¹. Such a concentration is an order of magnitude higher than the solution equilibrium concentration when N₂O in the shoot chamber is near ambient, so that movement of N₂O in the transpiration stream could have been significant. At least two experimental observations did not support this scenario. First, shoot N₂O emission during exposure to ¹⁵NO was not correlated with root zone N₂O production (Fig. 1B). Second, leaf N₂O flux fell below detectable limits during exposure to ¹⁵NH despite high rates of N₂O production by the rhizosphere (Table 1). The production of N₂O during ¹⁵NH exposure comes from the activity of nitrifying bacteria on the root surface (37), an activity we were able to completely shut down by using the nitrification inhibitor nitrapyrin. Thus, the absence of a correlation between shoot N₂O emission and rhizosphere N₂O production provides further evidence that photoassimilation of NO was the major source of N₂O emitted from leaves in this investigation. For investigations where N₂O transpiration was responsible for leaf N₂O flux (6), the concentrations of N₂O in the soil solution were nearly four orders of magnitude higher, at ≈326 μmol⋅liter⁻¹, than N₂O concentrations observed in our nutrient solutions. This may help to explain why N₂O transpiration was not a factor in this investigation.

The isotopic composition of N₂O emitted from leaves provided further verification that shoot NO assimilation was largely responsible for the observed N₂O flux. In the course of the experiments, we used nitrogen sources in nutrient solutions highly enriched in ¹⁵N (>99.6 atom % ¹⁵N as NH or NO). Thus, any N₂O produced by microbial metabolism in the rhizosphere, nitrification or denitrification, and subsequently transpired, would be predominantly composed of the ¹⁵N¹⁵N¹⁶O or ⁴⁶N₂O isoform. We grew our wheat plants in a nutrient solution with 200 μM NH plus 200 μM NO containing ¹⁵N close to natural abundance levels (0.3674% ¹⁵NH and 0.3665% ¹⁵NO, or > 99.6 atom % ¹⁴N). These plants stored relatively large quantities of NO in their tissues (8.76 ± 2.23 mM, mean ± SE, n = 6), of which 2.25 ± 1.65 mM was withdrawn and assimilated during the first 24-h exposure to ¹⁵NH. N₂O emitted during the assimilation of this internal NO would not be detected during exposure of roots to ¹⁵NH, because the isotopic signature on such stored NO was very similar to that of N₂O in the background air. However, if stored NO were being mobilized and assimilated in leaves during the subsequent period of root exposure to ¹⁵NO, or chloroplast reactants were participating in N₂O production, then some of the N₂O emitted would contain the ⁴⁵N₂O isoforms of ¹⁵N¹⁴N¹⁶O and ¹⁴N¹⁵N¹⁶O. Indeed, ⁴⁵N₂O was detected in the N₂O stream emitted from wheat leaves during exposure to ¹⁵NO. From the mass ratios of ⁴⁴N₂O, ⁴⁵N₂O, and ⁴⁶N₂O collected during such experiments, we estimated (30) that 5.19 ± 0.92 pmol⋅m⁻²⋅s⁻¹ (mean ± SE, n = 13), or about 20%, of the N₂O-N emitted by leaves came from the ¹⁴N stored in the plant. It is unlikely that bacterial mineralization of organic-N in the rhizosphere followed by nitrification and denitrification contributed to the ⁴⁵N₂O emitted. Roots treated with antibiotic cocktails (38) and nitrapyrin to eliminate rhizosphere microbial nitrogen transformations still emitted detectable quantities of ⁴⁵N₂O. Nonetheless, the largest fraction of the observed leaf N₂O flux was clearly generated by plant NO assimilation and not from N₂O produced by microbial processes in the rhizosphere.

The two enzymes responsible for plant NO assimilation, NO reductase (NR) and NO reductase (NiR), are located in the cytoplasm and chloroplasts, respectively. To determine whether NO or NO reduction was responsible for the N₂O emitted, we extracted NR, NiR, and intact chloroplasts from wheat leaves and assayed for N₂O production in vitro as described. Whereas NR assays did not produce any detectable N₂O, NiR assays and intact chloroplasts did (Fig. 2). These experiments provide in vitro evidence that the N₂O emitted from wheat leaves was generated by NO reduction to NH₃ in the chloroplasts, where NO undergoes transformation through the +2 to +1 oxidation state, as predicted.

Figure 2.

N₂O production in vitro from (A) intact chloroplasts (nmol N₂O-N g⁻¹ chlorophyll s⁻¹), (B) nitrate reductase (nmol N₂O-N g⁻¹ protein s⁻¹), and (C) nitrite reductase (nmol N₂O-N g⁻¹ protein s⁻¹). Intact chloroplasts, nitrate reductase, and nitrite reductase were extracted from fully expanded leaves of 2- to 3-week-old wheat plants (T. aestivum L. cv. Veery 10).

The leaf N₂O emissions we measured were small (Table 1), and it is not currently known how important fluxes of this magnitude are in relation to the global N₂O budget. Some recent micrometeorological N₂O flux measurements over grass swards failed to detect a higher N₂O emission rate than that observed by using static chambers over soil alone (39). However, the error in such measurements was too large, with coefficients of variation for the chamber data ranging from 0.42 to 1.83 (40, 41), to resolve plant N₂O emissions of the size measured in this investigation. In support of a role for plant N₂O emission, Hutchinson and Mosier (42) found that aerodynamic flux measurements over an irrigated corn field were always higher than, although never exceeding twice the mean of, fluxes measured by using soil chambers. Some recent chamber studies found that ryegrass canopies were responsible for 21.1% of the total N₂O emissions (4), and that Linum perenne canopies contributed to as much as 50% of N₂O flux to the atmosphere (43). Our results indicated that 0.02–0.2% of the NO-N assimilated by wheat was released as N₂O-N. Current estimates are that terrestrial land plants assimilate 1,200 Tg of N annually (44). Nearly half of this N is thought to be absorbed and assimilated as NO (45), of which 25–75% is assimilated in leaves (46). The NR and NiR enzymes involved are highly conserved among higher plants (47). Thus, if other terrestrial land plants behave as wheat does, we calculate that NO photoassimilation alone could produce from 0.03 to 0.9 Tg N₂O-N yr⁻¹. The uncertainty in the IPCC's estimates of global N₂O source emissions is substantial, with nearly two orders of magnitude in the range of estimated emission of N₂O from soils [0.6–14.8 Tg N y⁻¹ (1)]. Robertson et al. (25) tracked N₂O emissions for 9 years from soils of six cropping systems, including a nitrogen-intensive corn rotation and four successional communities, and found emissions to be at the lower end of the IPCC's calculated emission factor, which is based on fertilizer application. Our estimates of plant N₂O emissions represent ≈5–6% of the total amount of N₂O-N thought to be emitted by agricultural plant–soil systems alone (1, 44). These approximations do not include the quantity of N₂O that might be conducted to the atmosphere via the plant transpiration stream. Thus, our results suggest that higher plants could play an intriguing role in N₂O exchange not previously considered in biosphere–atmosphere interactions.

Abbreviation

IPCC

Intergovernmental Panel on Climate Change