Nitrate-Dependent N₂O Emission from Intact Soybean Nodules via Denitrification by Bradyrhizobium japonicum Bacteroids

Abstract

In the presence of nitrate, N₂O emission increased markedly from soybean roots inoculated with nosZ mutant of Bradyrhizobium japonicum, but not from soybean roots inoculated with a napA nosZ double mutant, indicating that B. japonicum bacteroids in soybean nodules are able to convert the exogenously supplied nitrate into N₂O via a denitrification pathway.

TEXT

Nitrous oxide (N₂O) is a key atmospheric greenhouse effect gas that not only affects global warming but also leads to destruction of the stratospheric ozone layer (15). Agricultural land is one of the major sources of N₂O (8, 12, 23). In particular, more N₂O is emitted from agricultural fields with legume crops than from fields with nonlegume crops (5, 11). Kim et al. (10) found that N₂O emission from fields with nodulating soybean was several times higher than that from fields with nonnodulating soybean at a flowering stage of soybean growth in the field, suggesting that nodulation enhanced N₂O emission. On the other hand, Sameshima-Saito et al. (17) reported that soybean nodules can take up a low concentration of N₂O from outside the nodules, equivalent to the natural concentration of N₂O in the air (approximately 0.31 ppm) and that this uptake depends on the nosZ gene, which encodes N₂O reductase in Bradyrhizobium japonicum. Therefore, the N₂O metabolism of nodulated soybean roots is complex, and the mechanism underlying N₂O emission from soybean fields has not yet been identified.

Under field conditions, N fertilization generally increases the nitrate concentration in the soil solution (7). Thus, we hypothesized that the increased nitrate supply may lead to an increase in N₂O emission from intact soybean root systems via a denitrification pathway in Bradyrhizobium japonicum. To test this hypothesis, we constructed a napA nosZ double mutant of B. japonicum USDA110; the wild-type (WT) versions of these genes encode dissimilatory nitrate reductase (3) and N₂O reductase (17), respectively.

The bacterial strains and plasmids we used are listed in Table 1. Bradyrhizobium cells were grown at 30°C in HM salt medium supplemented with 0.1% arabinose and 0.025% (wt/vol) yeast extract (Difco, Detroit, MI) (18). HM medium was further supplemented with trace metals (HMM medium) for the denitrification assay (17). Escherichia coli cells were grown at 37°C in Luria-Bertani medium (16). Antibiotics were added to the media: tetracycline (Tc) at 100 μg/ml, spectinomycin (Sp) at 100 μg/ml, streptomycin (Sm) at 100 μg/ml, kanamycin (Km) at 100 μg/ml, and polymyxin B at 50 μg/ml for B. japonicum and Tc at 15 μg/ml, Sp at 50 μg/ml, Sm at 50 μg/ml, Km at 50 μg/ml, and ampicillin (Ap) at 100 μg/ml for E. coli.

Table 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristic(s)a	Source or reference
Strains
Bradyrhizobium japonicum
USDA110	Wild type, nosZ+	9
USDA110 ΔnosZ::Tc	USDA110 derivative, nosZ::del/ins Tc cassette; Tcr	This study
USDA110 ΔnapA	USDA110 derivative, napA::Ω cassette; Spr Smr	This study
USDA110 ΔnapA ΔnosZ	USDA110 derivative, nosZ::del/ins Tc cassette Tcr; napA::Ω cassette Spr Smr	This study
Escherichia coli
DH5α	recA, cloning strain	Toyobo Inc.
Plasmids
pK18mob	Cloning vector, pMB1ori oriT Kmr	21
brp16565	pUC18 carrying a 10.8-kb nap gene cluster	9
pRK2013	ColE1 replicon carrying RK2 transfer gene; Kmr	22
p34S-Tc	Plasmid carrying 2.0-kb Tcr cassette; Tcr Apr	4
pHP45Ω	Plasmid carrying 2.1-kb Ω cassette; Spr Smr Apr	14
pK18mob-nap	pK18mob carrying 7.5-kb-HindIII fragment including nap gene cluster; Kmr	This study
pK18mob-ΔnapA::Ω	pK18mob carrying 9.5-kb fragment including nap gene cluster and Ω cassette; Kmr	This study
pTZ18RnosZ	pTZ18R carrying 4.0-kb-nosZ fragment; Apr	17
pK18mob-ΔnosZ::Tc	pK18mob carrying 4.6-kb fragment including nosZ and Tc cassette; Kmr Tcr	This study

^aThe resistance genes included the genes coding for resistance to ampicillin (Ap^r), kanamycin (Km^r), polymyxin B, streptomycin (Sm^r), spectinomycin (Sp^r), and tetracycline (Tc^r).

Isolation of plasmids, DNA ligation, and transformation of E. coli were performed as described by Sambrook et al. (16). DNA preparation and Southern hybridization were carried out as described previously (20).

To construct a napA nosZ double mutant with different antibiotic markers, we reconstructed a nosZ single-deletion mutant in B. japonicum USDA110 using a Tc cassette. pK18mob-ΔnosZ::Tc was constructed using the Tc cassette from the previously created plasmid pTZ18RnosZ (17) and was subjected to triparental mating with a helper plasmid (pRK2013), resulting in USDA110 ΔnosZ::Tc (Table 1). For the napA mutation, a 7.5-kb DNA fragment carrying napEDABC was excised from HindIII-digested brp16565 (9) and inserted into the HindIII site of pK18mob (Fig. 1). A 2.0-kb Ω cassette derived from pHP45 (14) was then inserted at a BstPI site within napA of B. japonicum USDA110 (Fig. 1), resulting in pK18mob-ΔnapA::Ω (Fig. 1). Triparental mating with wild-type USDA110 and USDA110 ΔnosZ::Tc as recipients was conducted on HMM agar plates using pK18mob-ΔnapA::Ω and pRK2013, resulting in USDA110 ΔnapA and USDA110 ΔnapA ΔnosZ, respectively (Table 1). Double-crossover events were verified by Southern hybridization.

Fig. 1.

Construction of the napA insertion mutant. Cloned fragments in pK18mob-nap and pK18mob-ΔnapA::Ω are shown alongside the physical map of the nap gene cluster of B. japonicum USDA110. H, HindIII site; B, BstPI site. The 7.5-kb HindIII fragment containing the napEDABC genes was cloned into pK18mob, resulting in pK18mob-nap. After BstPI digestion and the blunt-end reaction, the Ω cassette was inserted into the BstPI site in napA, resulting in pK18mob-ΔnapA::Ω. By triparental mating using pK18mob-ΔnapA::Ω as a donor, USDA110 ΔnapA and USDA110 ΔnapA ΔnosZ were constructed from wild-type USDA110 and USDA110 ΔnosZ::Tc, respectively (see the text for details).

To confirm the denitrification capability of the mutants in culture, we measured the end products (N₂O and N₂) and the substrate (nitrate), as described previously (18, 19). Briefly, the aerobically grown cells (70 μl) were anaerobically incubated in HMM medium supplemented with 2 mM ¹⁵N-KNO₃ (99.6 atom%) at 30°C for 7 days in a test tube (15 ml of broth with a 19-ml gas phase). The ¹⁵N₂ and ¹⁵N₂O in the gas phase and ¹⁵N-nitrate in the liquid phase were determined after incubation for 7 days.

We germinated surface-sterilized soybean seeds (Glycine max cv. Enrei), and transplanted them into Leonard jars (17). We used cell suspensions of the wild-type and mutant strains of B. japonicum USDA110 (1 ml) for inoculation at 1 × 10⁷ cells per seed. Plants were grown for 25 days in an LH200 growth chamber at 25°C with a 16-h light/8-h dark cycle (Nippon Medical & Chemical Industries, Tokyo, Japan). A nitrogen-free nutrient solution was periodically supplied to the jars (13). The whole root system of the soybean plant was inserted into a 51-ml test tube (19.4 mm in diameter by 176 mm in height) containing 20 ml of nitrogen-free nutrient solution (13) in the presence or absence of 5 mM potassium nitrate (Fig. 2). The test tube was then sealed with a silicone rubber stopper equipped with a gas sampling port and a hole for the plant's stem that was sealed with silicon gum after insertion of the plant. This setup enclosed the nodulated root system of intact soybeans in the test tube, leaving the aerial parts in the open air (Fig. 2). The N₂O concentration of the gas phase in the chamber was monitored using a GC-14BPsE gas chromatograph (Shimadzu, Kyoto, Japan) equipped with an electron capture detector and tandem-packed columns of Porapak N and Porapak Q (Shimadzu), as described previously (17).

Fig. 2.

Closed-test-tube hydroponic system used for measurements of N₂O emission and uptake by the nodulated roots of intact soybean plants.

Free-living cells of the three mutants constructed in this work (Table 1) were examined in the denitrification assay (Table 2). The end product of wild-type USDA110 was N₂, whereas that of USDA110 ΔnosZ::Tc was N₂O. Both strains consumed all NO₃⁻. Thus, USDA110 ΔnosZ::Tc lost the ability to reduce N₂O as a result of Tc cassette insertion into the partially deleted nosZ gene (Table 2), which is similar to previous results with USDA110 ΔnosZ (17). USDA110 ΔnapA and USDA110 ΔnapA ΔnosZ did not consume NO₃⁻ and produced no N₂O and no N₂ (Table 2), indicating that they had lost the ability to reduce nitrate as a result of the Ω cassette insertion into napA (Fig. 1). Although cassette insertion often has a polar effect on downstream genes, such effect does not alter the interpretation because the downstream genes encode products for the same enzymatic processes (Fig. 1) (17).

Table 2.

Denitrification capability of B. japonicum mutants constructed in this work

Strain or condition	Amt of end product (μmol/15 ml culture)a
	NO3−	N2O	N2
USDA110 (WT)	<0.1	<0.1	32.2
USDA110 ΔnosZ::Tc	<0.1	28.0	<0.1
USDA110 ΔnapA	28.0	<0.1	<0.1
USDA110 ΔnapA ΔnosZ	28.7	<0.1	<0.1
Uninoculatedb	28.5	<0.1	<0.1

^aGray shading indicates the major end product (N₂O and N₂) or a high level of substrate (NO₃⁻). Values are expressed as means of duplicate determinations.

^b“Uninoculated” refers to medium that was not inoculated with B. japonicum cells.

After soybean plants inoculated with wild-type USDA110 (WT), USDA110ΔnosZ::Tc (ΔnosZ), or USDA110 ΔnapA ΔnosZ (ΔnapA ΔnosZ) were established in the closed-root-system chambers containing air in the absence or presence of 5 mM potassium nitrate in hydroponic solution (Fig. 2), the N₂O concentration was periodically monitored (Fig. 3). In the absence of nitrate (Fig. 3A), the soybean roots nodulated by the ΔnosZ and ΔnapA ΔnosZ strains showed no change in the N₂O concentration in air, whereas the roots nodulated by wild-type strain USDA110 showed a continuous decrease in the N₂O concentration, as was reported previously (17). In the presence of nitrate (Fig. 3B), the roots nodulated by the ΔnosZ strain emitted N₂O, whereas the roots nodulated by the ΔnapA ΔnosZ mutant showed no change in the N₂O concentration. Both mutants (ΔnosZ and ΔnapA ΔnosZ) differed only in whether they possessed the napA mutation in a common nosZ background (Table 1), and their differences in N₂O emission from the nodulated roots were statistically significant (Fig. 3B). Therefore, we conclude that the wild-type napA gene is required for N₂O emission from soybean roots nodulated by B. japonicum lacking the wild-type nosZ gene in the presence of nitrate.

Fig. 3.

N₂O emission and uptake in the closed-root-chamber system (Fig. 2) by soybean nodules formed with the wild-type (WT), ΔnosZ mutant, or ΔnapA ΔnosZ double mutant of Bradyrhizobium japonicum USDA110 in the absence (A) and presence (B) of nitrate (5 mM KNO₃). Asterisks indicate a significant difference in the N₂O concentration from that of the napA nosZ double mutant at P < 0.05 (*) and P < 0.01 (**) (t test, triplicate determination). Uninoculated soybean roots showed no emission and no uptake of N₂O (data not shown), which was identical to the N₂O profile of soybeans inoculated with the ΔnapA ΔnosZ double mutant.

The napA gene encodes nitrate reductase, which catalyzes the first step in denitrification (NO₃⁻ → NO₂⁻ → NO → N₂O → N₂) in B. japonicum USDA110 (Fig. 4), and free-living cells of the napA mutant lost the capability to reduce nitrate (Table 2). Our results strongly suggest that endosymbiotic cells of B. japonicum that lack the wild-type nosZ gene in soybean nodules can convert exogenously supplied nitrate into N₂O via their denitrification pathway (Fig. 4).

Fig. 4.

Schematic representation of nitrate-dependent N₂O emission from soybean nodules formed by the wild-type (WT) strain, the ΔnosZ mutant, and the ΔnapA ΔnosZ double mutant of Bradyrhizobium japonicum USDA110. napA, nirK, norCB, and nosZ are structural genes coding for periplasmic nitrate reductase, nitrite reductase, nitric oxide reductase, and nitrous oxide reductase, respectively, in B. japonicum USDA110 (1).

Natural nosZ-negative populations of B. japonicum, which lack nosZ and N₂O reductase activity, are often dominant in soybean fields (19, 20). Therefore, if nosZ-negative populations of B. japonicum are dominant in soybean fields and if nitrate concentrations in the soil solution increase as a result of high levels of N fertilization or high soil N availability, N₂O could be emitted from the nodulated soybean roots. Indeed, the variable in field experiments that best explained cumulative N₂O emission during the whole soybean growing season was the soil nitrate level (2).

It is very surprising that root systems nodulated by wild-type USDA110 decreased the atmospheric N₂O concentration (approximately 0.3 ppm) in the presence of nitrate (Fig. 3B). This result suggests that N₂O reductase activity overcame any N₂O production that resulted from the added nitrate (Fig. 4). Therefore, these results suggest that there is high potential to reduce N₂O emissions by nosZ-carrying strains of B. japonicum to mitigate N₂O emission from soybean rhizosphere (17). Recently, N₂O reduction by nosZ-carrying inoculants was shown in soil-filled pots planted with soybean (6).