Generation of Bradyrhizobium japonicum Mutants with Increased N₂O Reductase Activity by Selection after Introduction of a Mutated dnaQ Gene

Abstract

We obtained two beneficial mutants of Bradyrhizobium japonicum USDA110 with increased nitrous oxide (N₂O) reductase (N₂OR) activity by introducing a plasmid containing a mutated B. japonicum dnaQ gene (pKQ2) and performing enrichment culture under selection pressure for N₂O respiration. Mutation of dnaQ, which encodes the epsilon subunit of DNA polymerase III, gives a strong mutator phenotype in Escherichia coli. pKQ2 introduction into B. japonicum USDA110 increased the frequency of occurrence of colonies spontaneously resistant to kanamycin. A series of repeated cultivations of USDA110 with and without pKQ2 was conducted in anaerobic conditions under 5% (vol/vol) or 20% (vol/vol) N₂O atmosphere. At the 10th cultivation cycle, cell populations of USDA110(pKQ2) showed higher N₂OR activity than the wild-type strains. Four bacterial mutants lacking pKQ2 obtained by plant passage showed 7 to 12 times the N₂OR activity of the wild-type USDA110. Although two mutants had a weak or null fix phenotype for symbiotic nitrogen fixation, the remaining two (5M09 and 5M14) had the same symbiotic nitrogen fixation ability and heterotrophic growth in culture as wild-type USDA110.

Nitrous oxide (N₂O) is a key atmospheric greenhouse gas that contributes to global climate change through radiative warming and depletion of stratospheric ozone (4, 19). Agricultural field soil is a major source of N₂O through the microbial transformation of nitrogen in soil (10, 14, 24) and contributes significantly to the net increase in atmospheric N₂O (4, 25, 27, 36). Soybean fields are also sources of emission of N₂O, probably because of the presence of their nodulated roots in the rhizosphere (18, 40). Several attempts have been made to reduce the emission of N₂O from agricultural systems (24, 25). The use of slow- or controlled-release fertilizers or nitrification inhibitors chemically reduces the emission of N₂O from agricultural systems (2, 36). However, to our knowledge, microorganisms have not been used to biologically reduce emissions of N₂O from agricultural systems.

The complete denitrification of nitrate to dinitrogen (N₂) by bacteria is generally an anaerobic respiratory process wherein the last step is mediated by N₂O reductase (N₂OR) (43, 44). The corresponding structural gene is nosZ, which is assembled in the nosRZDFYL gene operon (43, 44). Bradyrhizobium japonicum, a symbiotic nitrogen-fixing soil bacterium associated with soybeans (7, 15), is capable of denitrification (32, 33, 39). The B. japonicum nosZ gene encodes N₂OR, which mediates the reduction of N₂O to N₂ (31, 39). Sameshima-Saito et al. (31) found that the introduction of cosmids carrying nosRZDFYL into B. japonicum markedly increases N₂OR activity, suggesting that there is room for the enhancement of N₂OR activity in B. japonicum.

One molecular breeding procedure in microbes is the use of proofreading-deficient DNA polymerase mutators, which are cells that have higher mutation rates than the wild type (22). In Escherichia coli, the epsilon subunit of DNA polymerase III supplies exonuclease activity and is encoded by dnaQ (mutD) (22). Mutation of the dnaQ gene dominantly impairs proofreading activity, resulting in a very strong mutator phenotype via unedited replication errors during growth (9). The availability of proofreading-deficient mutators for the isolation of spontaneous mutants under selective pressure has been reported in antibiotic-resistant mutants of E. coli (37) and thermotolerant mutants of Saccharomyces cerevisiae (34).

In this work, we aimed to obtain beneficial mutants of B. japonicum USDA110 with increased N₂OR activity by introducing a plasmid containing a mutated B. japonicum dnaQ gene (pKQ2) and then performing enrichment culture under selection pressure for N₂O respiration. Examination of the genome sequence of B. japonicum USDA110 has indicated that the blr0640 gene is a dnaQ homolog (15). We therefore constructed a pKQ2 plasmid containing an artificial mutation of blr0640.

MATERIALS AND METHODS

Bacterial strains and media.

The bacterial strains and plasmids used in this study are listed in Table 1. Bradyrhizobium japonicum cells were grown at 30°C in HM salt medium (6) supplemented with 0.1% arabinose and 0.025% (wt/vol) yeast extract (Difco, Detroit, MI). HM medium was further supplemented with trace metals (HMM medium) for the denitrification assay (32). HM liquid media were cultured with reciprocal shaking at 120 rpm (flask) and 300 rpm (test tube). E. coli cells were grown at 37°C in Luria-Bertani medium (29). Antibiotics were added to the media in the following concentrations: for B. japonicum, 50 or 100 μg tetracycline (Tc)/ml, 50 μg polymyxin B/ml, and 100 μg kanamycin (Km)/ml; for E. coli, 10 μg Tc/ml and 50 μg Km/ml.

TABLE 1.

Bacterial strains and cosmids used

Strain or plasmid	Relevant characteristicsa	Source or reference
Strains
Bradyrhizobium japonicum
USDA110	Wild type; Tcs Nos+ Nod+ Fix+	15
S2	USDA110 higher-salt-tolerance mutant; Tcs	This study
S4	USDA110 higher-salt-tolerance mutant; Tcs	This study
S5	USDA110 higher-salt-tolerance mutant; Tcs	This study
S9	USDA110 higher-salt-tolerance mutant; Tcs	This study
5M08	USDA110 higher-N2OR-activity mutant; Tcs Nos++ Nod+ Fix−	This study
5M09	USDA110 higher-N2OR-activity mutant; Tcs Nos++ Nod+ Fix+	This study
5M14	USDA110 higher-N2OR-activity mutant; Tcs Nos++ Nod+ Fix+	This study
20M19	USDA110 higher-N2OR-activity mutant; Tcs Nos++ Nod+ Fix+/−	This study
Escherichia coli DH5α	Cloning strain; recA	Toyobob
Plasmids
pKS800	Derivative of broad-host-range cosmid pLAFR1; IncP Tcr	11
pKQ1	pKS800 carrying aph promoter and wild-type blr0640; Tcr	This study
pKQ2	pKS800 carrying aph promoter and two-base-pair substitution mutant of blr0640; Tcr	This study
pRK2013	ColE1 replicon carrying RK2 transfer genes; Kmr	8
pUC4-KIXX	Plasmid carrying 1.6-kb aph cassette; Apr Kmr	Amersham-Pharmaciac

^aNos, N₂OR activity; Nos phenotype is expressed as + (N₂OR activity of USDA110) or ++ (N₂OR activity significantly higher than that of USDA110). Fix phenotype (nitrogen fixation) is expressed as + (nitrogen fixation activity similar to that of USDA110), +/− (nitrogen fixation activity significantly lower than that of USDA110), or − (no nitrogen fixation activity). Nod⁺, positive for nodule formation on soybeans. blr0640, DNA polymerase III epsilon chain gene homolog; aph, aminoglycosidase-3-O-phosphotransferase gene.

^bToyobo, Inc., Tokyo, Japan.

^cAmersham-Pharmacia Biotech, Uppsala, Sweden.

Construction of the mutator plasmid.

DNA fragments of wild-type and mutated dnaQ genes, dnaQ^WT and dnaQ^exo-1, were amplified with two forward primers (dnaQ^WT-F, 5′-CCGCTCGAGATGCGCGAAATCGTTCTCGACACCGAAACC, and dnaQ^exo-1-F, 5′-CCGCTCGAGATGCGCGAAATCGTTCTCGCCACCGCAACC) and a reverse primer (dnaQ-R, 5′-CGCGGATCCCTAACCAGCCGGATTGACGGTAACTG) containing the XhoI and BamHI sites (underlined) from the B. japonicum genome. These fragments were digested with XhoI and BamHI and were introduced into the same sites of the cosmid vector pKS800. For dnaQ gene expression, the aph promoter fragment (U00004) was amplified with a primer pair (aphP-F, 5′-TACATCCTCGAGAAGCCAGTCCGCAGAAACGGTGC, and aphP-R, 5′-TGATGCCTCGAGGATCCTCATCCTGTCTCTTGATCAGA) containing the XhoI site (underlined) from pUC4-KIXX (Amersham-Pharmacia Biotech, Uppsala, Sweden). This fragment was digested with XhoI and introduced into the same site in the wild-type and mutated dnaQ plasmids. Finally, we constructed two plasmids named pKQ1 and pKQ2 carrying the wild-type and mutated dnaQ genes, respectively (Table 1 and Fig. 1). These two plasmids and the vector pKS800 were transformed into E. coli DH5α, and then each was introduced into B. japonicum USDA110 by means of triparental mating, using pRK2013 as a helper plasmid, on HM agar plates (26, 31, 35, 41). Transconjugants were selected on HM agar plates containing 100 μg Tc/ml and 50 μg polymyxin B/ml.

FIG. 1.

Construction of plasmids carrying wild-type dnaQ (pKQ1) or mutated dnaQ (pKQ2). Asterisked nucleotides indicate positions of the two-base-pair substitutions. Boxes indicate wild-type and mutated codons in B. japonicum dnaQ homolog for conversion of ₇Asp (GAC) and ₉Glu (GAA) into ₇Ala (GCC) and ₉Ala (GCA), respectively. aphP is the promoter of the Km resistance gene from pUC4-KIXX (Table 1). Arrows with prefix “P” followed by 3W, 3M, 1, 2, 4, 5, or 6 show the positions and directions of PCR primers. DNA sequences of primers 3W and 3M are underlined. “MCS” indicates multicloning sites for pKS800 (Table 1). The resulting plasmids (pKQ1 and pKQ2) were 22.4 kb. X, XhoI; B, BamHI.

Antibiotic resistance.

Bradyrhizobium japonicum USDA110 cells were cultured at 30°C for 7 days until they had reached a concentration of 10⁹ cells/ml. The cultures were then serially diluted with HM broth and inoculated on HM agar plates with and without 100 μg Km/ml. After incubation of the plates at 30°C for 7 days, the frequencies of occurrence of Km-resistant colonies were calculated from CFU on HM agar plates with and without 100 μg Km/ml.

Mean generation time.

Precultures (50 μl; 10⁹ cells/ml) of USDA110 cells were inoculated into 5 ml of HM broth. The turbidity (A₆₆₀) of cultures grown aerobically at 30°C was measured every 24 h with a UV-1200 spectrophotometer (Shimadzu, Kyoto, Japan). The mean generation time was calculated as described previously (30).

Selection for salt stress.

USDA110 carrying pKS800, pKQ1, or pKQ2 (1 ml; 10⁹ cells/ml) was inoculated into 100 ml of HM broth in a flask supplemented with 50 μg Tc/ml and 75 mM NaCl and then grown at 30°C for 10 days as the first cultivation. The resultant cultures (1 ml) were reinoculated into 100 ml HM broth medium with 75 mM NaCl and 50 μg Tc/ml and then incubated at 30°C for 14 days. Periodic transfer (1 ml) was repeated at intervals of 10 to 14 days up to the seventh cultivation cycle. To assay the increase of salt tolerance and the selection for increased NaCl tolerance, cultures (1 ml) were also inoculated into 100 ml HM broth medium with 100 mM NaCl and 50 μg Tc/ml at the first and seventh cultivation cycles.

Selection for N₂O respiration.

USDA110 carrying pKS800 or pKQ2 (200 μl; 10⁹ cells/ml) was inoculated into 5 ml of HM broth medium with 50 μg Tc/ml. To achieve anaerobic N₂O respiration conditions, N₂O gas was introduced at a final concentration of 5 or 20% (vol/vol) (N₂ balance) into the headspace (29 ml) of a test tube (18-mm diameter by 180-mm height) sealed with a butyl rubber stopper. These cultures were incubated at 30°C for 9 days as a first cultivation. The resultant cultures (50 μl) were reinoculated into 5 ml HM broth medium with 50 μg Tc/ml and then grown for 7 days under conditions of anaerobic N₂O respiration as a 2nd cultivation cycle; periodic transfer (50 μl) was repeated at intervals of 7 to 14 days up to the 10th cultivation cycle.

N₂OR activity.

B. japonicum cells were washed with HMM broth by centrifugation (5,000 × g, 15 min, 4°C) and were suspended at 10⁹ cells/ml in the broth. N₂OR activity was determined by using a GC-17A gas chromatograph (Shimadzu) equipped with a ⁶³Ni electron capture detector and a CP-PoraBOND Q-capillary column (internal diameter, 0.32 mm; length, 25 m; Varian, Palo Alto, CA), as described previously (31).

Plant inoculation.

Surface-sterilized soybean seeds (Glycine max ‘Enrei’) were germinated in sterile vermiculite for 2 days at 25°C and then transplanted to a Leonard jar pot (20, 38, 42) which contained sterile vermiculite and nitrogen-free nutrient solution (23, 26). The seeds were then inoculated with Bradyrhizobium japonicum at 1 × 10⁷ cells per seed. Plants were grown in a growth chamber (LH200; Nippon Medical & Chemical Industries, Tokyo, Japan) for 4 weeks at 25°C with 16 h light and 8 h dark.

Single-colony isolation via plant passage.

After soybean cultivation for 4 weeks, nodules excised from the soybean roots were sterilized with NaClO solution (0.5% wt/vol) and washed 10 times with sterilized water. Each sterilized nodule was cut in half, and the bacteroid cells were streaked onto an HM agar plate for isolation of single colonies. After incubation at 30°C for 1 week, single colonies from each nodule were inoculated onto HM agar plates with or without 100 μg Tc/ml to ensure plasmid curing.

Symbiotic phenotype.

After soybean cultivation for 4 weeks, the nodulated roots of one plant were introduced into 125-ml vials. Acetylene gas was injected at a final concentration of 10% (vol/vol), and the roots were incubated for 20 min at 20°C. The ethylene concentration was determined with a Shimadzu GC-18A gas chromatograph (Shimadzu) equipped with a flame ionization detector and a Porapack N column (28). Then, the number and weight of the nodules per plant were determined.

PCR verification of plasmid elimination.

Total DNA was prepared with an AquaPure genomic DNA isolation kit (Bio-Rad Laboratories, Hercules, CA) as a PCR template. Four primer pairs were designed for PCR amplification of plasmids carrying the dnaQ of the mutant or wild type (Fig. 1). The primer sequences were primer 1, 5′-TAGACTGGGCGGTTTTATGG; primer 2, 5′-ATCTCGACACAGCCGATTTC; primer 4, 5′-CTCGGCATTGATGAAGCTGA-3′; primer 5, 5′-ATCGACATCCTCAACGGAAG; and primer 6, 5′-CTTTCCCGCTCACTTTTCAG. Primers 3W and 3M are shown in Fig. 1. ExTaq polymerase (Takara, Osaka, Japan) was used for PCR amplification. For PCR primers 1, 2, 5, and 6, the PCR cycles were 1 cycle of 94°C for 5 min, 30 cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 30 s, and a final extension cycle of 72°C for 10 min. For PCR primers 3W, 3M, and 4, the PCR cycles were 1 cycle of 94°C for 5 min, 25 cycles of 94°C for 30 s, 65°C for 30 s, and 72°C for 30 s, and a final extension cycle of 72°C for 10 min.

RESULTS

Construction of the B. japonicum USDA110 mutator.

₇Asp (GAC) and ₉Glu (GAA) of blr0640 are amino acid sequence motifs that are conserved among gram-negative bacteria and the exonuclease I motif of the dnaQ gene of E. coli (9, 13). Therefore, we constructed a two-base-pair substitution mutant of blr0640 by replacing ₇Asp (GAC) with ₇Ala (GCC), and ₉Glu (GAA) with ₉Ala (GCA) by PCR (Fig. 1). The mutated dnaQ gene and aph promoter fragment were inserted into the BamHI/XhoI sites of pKS800 (11), resulting in pKQ2 (Fig. 1). pKQ2 was introduced into B. japonicum USDA110 by transconjugation. pKS800 and pKS800 carrying the wild-type blr0640(pKQ1) were also introduced into B. japonicum USDA110 as controls (Fig. 1).

The generation time of USDA110 carrying pKQ2 was similar to that of USDA110 carrying pKS800 or pKQ1 (Table 2). In contrast, the frequency of occurrence of the spontaneous Km-resistant mutant of USDA110 carrying pKQ2 was significantly higher than that of USDA110 carrying pKS800 or pKQ1 (Table 2), indicating that pKQ2 conferred in trans a mutator phenotype on B. japonicum USDA110.

TABLE 2.

Generation time and frequency of occurrence of spontaneous Km resistance in B. japonicum USDA110 carrying pKS800, pKQ1, or pKQ2^a

Strain	Generation time (h)b	Frequency of occurrence of Kmr colonies (ratio relative to value for pKS800)
USDA110(pKS800)	15.1 ± 0.2	1.6 × 10−7 ± 0.1 × 10−7 (1)*
USDA110(pKQ1)	14.9 ± 0.7	1.5 × 10−7 ± 0.2 × 10−7 (1)*
USDA110(pKQ2)	15.1 ± 0.5	1.5 × 10−5 ± 0.2 × 10−5 (95)#

^aValues are expressed as means ± standard deviations (n = 3). Entries followed by the same symbol (* or #) do not differ significantly by t test (P < 0.01).

^bGeneration times were calculated from growth in HM broth with 50 μg of Tc/ml.

Isolation of salt-tolerant mutants.

To examine whether the introduction of pKQ2 would produce B. japonicum mutants of interest under selective pressure, we tried to isolate mutants with increased salt tolerance. Although USDA110 was unable to grow in medium supplemented with 100 mM NaCl (5), 75 mM NaCl permitted its slow growth (data not shown). Thus, USDA110 cells carrying pKQ2 (as mutator cells) and USDA110 cells carrying pKS800 or pKQ1 (as control cells) were cultured repeatedly in HM medium containing 75 mM NaCl (Fig. 2). The growth of mutator cells [USDA110(pKQ2)] was slower than that of control cells [USDA110(pKQ1)] in the first cultivation cycle (Fig. 3A, left). This was not surprising, because mutation of dnaQ generally reduces the growth rate (9, 22). In contrast, in the seventh cultivation cycle, the growth of the mutator cells was similar to that of the control cells (Fig. 3A, right). When we used stronger salt stress (100 mM NaCl), after the seventh cultivation cycle we observed vigorous growth of the mutator cell [USDA110(pKQ2)] population (Fig. 3B, right). In contrast, no growth was observed in control cells [USDA110(pKQ1)] even after the seventh cultivation cycle (Fig. 3B, right).

FIG. 2.

Strategies for isolation of Bradyrhizobium japonicum (Bj) mutants with increased NaCl tolerance and increased N₂OR activity by pKQ2 introduction and enrichment culture under selective pressures. Box shows steps by which an increased mutation rate was produced in the cell population of B. japonicum USDA110(pKQ2) (see text).

FIG. 3.

Growth of cell populations under salt-stress selection and of mutants of Bradyrhizobium japonicum USDA110 with increased salt resistance. (A) Cells were repeatedly cultivated in 75 mM NaCl to adapt to salt stress (Fig. 2). Cell growth curves of first and seventh culture cycles are shown. (B) Cells from first and seventh culture cycles were assayed to examine whether the salt tolerance of cell populations increased in 100 mM NaCl. (C) Growth of isolates from seventh culture cycle of USDA110(pKQ2) population in 100 mM NaCl (panel B, right). Error bars indicate standard deviations (n = 3). 110, strain USDA110; OD660, optical density at 660 m.

Mutator cells [USDA110(pKQ2)] from the seventh cultivation cycle were inoculated into soybeans (Fig. 2). Tc-sensitive single colonies from the soybean nodules were then isolated on HM agar plates. Four Tc-sensitive mutants, S2, S4, S5, and S9, were able to grow even in HM broth containing 100 mM NaCl, whereas no growth was observed in wild-type USDA110 (Fig. 3C). This result showed that the mutants isolated still had an increased-salt-tolerance phenotype, although they had probably lost pKQ2.

Isolation of mutants with increased N₂OR activity.

In light of the results of the salt-stress experiment, we designed a strategy for isolating mutants with increased N₂OR activity (Fig. 2). A series of cultivations of mutator and control cells were performed in anaerobic conditions under 5% (vol/vol) or 20% (vol/vol) N₂O. The growth of mutator cells [USDA110(pKQ2)] was slower than that of control cells [USDA110(pKS800)] during the first cultivation cycle (Fig. 4A and B, left). At the 10th cultivation cycle, the growth of the mutator cells [USDA110(pKQ2)] was faster than that of the control cells [USDA110(pKS800)] under 5% N₂O and similar to that of the control cells [USDA110(pKS800)] under 20% N₂O (Fig. 4A and B, right).

FIG. 4.

Growth of cell populations under N₂O respiration and of mutants of Bradyrhizobium japonicum USDA110 with increased N₂OR activity. (A and B) Growth of 1st and 10th culture cycles of USDA110(pKS800) and USDA110(pKQ2) populations in HM medium in the presence of 5% (A) and 20% (B) N₂O. (C) N₂OR activity of 10th culture cycles of USDA110(pKS800) and USDA110(pKQ2) populations. (D) N₂OR activity of isolates from 10th culture cycle of USDA110(pKQ2) population. Error bars indicate standard deviations. (n = 3). 110, strain USDA110; OD660, optical density at 660 m.

We therefore determined the N₂OR activity of cells from the 10th cultivation cycle. The mutator populations serially cultivated in 5% or 20% N₂O atmospheres took up N₂O more rapidly than the control populations (Fig. 4C). Based on the reduction rate with time (Fig. 4C), the N₂OR activities of mutator cell populations [USDA110(pKQ2)] from the 5 and 20% N₂O selections were 77 nmol h⁻¹ 10⁹ cells⁻¹ and 65 nmol h⁻¹ 10⁹ cells⁻¹, respectively. On the other hand, the N₂OR activity of the control cell populations was approximately 7 nmol/h per 10⁹ cells in USDA110(pKS800) from both the 1st and the 10th cultivation cycles under 5% or 20% N₂O respiration selection.

Mutator cells from the 10th cultivation cycle under selection pressure of N₂O respiration were inoculated into soybeans (Fig. 2). From this plant passage, we obtained 31 Tc-sensitive isolates; we measured their N₂OR activities as a first screening (data not shown) and selected four promising mutants which took up N₂O more rapidly than wild-type USDA110 (Fig. 4D). The N₂OR activities of 5M08, 5M09, 5M14, and 20M19 (85 to 138 nmol h⁻¹ 10⁹ cells⁻¹) were 11, 9, 7, and 9 times those of the wild-type USDA110 (12 nmol h⁻¹ 10⁹ cells⁻¹) (Fig. 4D and Table 3). This result showed that the mutants still had increased N₂OR activity; their activities were similar to that of the pKQ2 mutator population (Fig. 4C).

TABLE 3.

Free-living and symbiotic phenotypes of B. japonicum USDA110 mutants with increased N₂OR activity^a

Strain	N2OR activity (nmol h−1 109 cells−1)b	Generation time (h)c	Frequency of occurrence of Kmr colonies	Number of nodules (plant−1)	Nodule wt (mg [fresh wt] nodule−1)	Nitrogen-fixing activity (μmol C2H4 produced h−1 g of nodule [fresh wt]−1)d
USDA110	12 ± 1¶	10.0 ± 1.2*	1.9 × 10−7 ± 0.6 × 10−7*	52 ± 7*	13.0 ± 1.7*	12.2 ± 3.8*
5M08	107 ± 0#	9.6 ± 0.0*	1.3 × 10−7 ± 0.3 × 10−7*	102 ± 8#	5.8 ± 0.9#	0.0 ± 0.0#
5M09	85 ± 4§	9.7 ± 0.6*	1.1 × 10−7 ± 0.4 × 10−7*	52 ± 14*	11.3 ± 4.2*	14.4 ± 2.8*
5M14	114 ± 2#,§	10.1 ± 0.6*	2.0 × 10−7 ± 0.4 × 10−7*	48 ± 12*	10.4 ± 3.1*	26.8 ± 6.2*
20M19	138 ± 2*	11.7 ± 0.9*	3.3 × 10−7 ± 1.0 × 10−7*	49 ± 10*	13.5 ± 3.4*	0.8 ± 0.8#

^aValues are expressed as means ± standard deviations (n = 3). Values in the same column followed by the same symbols (*, #, §, and ¶) do not differ significantly by t test (P < 0.05).

^b5M08, 5M09, and 5M14 were obtained from 5% N₂O selection, while 20M19 was derived from 20% N₂O selection.

^cGeneration times were calculated from growth in HM broth medium.

^dNitrogen-fixing activity was evaluated by measuring ARA.

Verification of pKQ2 elimination.

The frequencies of occurrence of spontaneous Km-resistant mutants among the four mutants (20M19, 5M08, 5M09, and 5M14) were of the order of approximately 10⁻⁷, the same level as in the wild-type USDA110 (Table 3) and lower than that of USDA110(pKQ2) (Table 2). In addition, PCR fragments were not detected from the four mutants when P1/P2, P3M/P4, and P5/P6 were used as PCR primers, expect for P3W/P4 (see Fig. S1 in the supplemental material). On the other hand, all PCR fragments were detected from USDA110(pKQ2) (see Fig. S1 in the supplemental material). These results indicate that pKQ2 was genetically and physically eliminated in the mutants.

Growth and symbiotic traits of the mutants isolated.

The generation times of the four mutants in HM broth were not significantly different from that of the wild-type USDA110 (Table 3). Mutant 5M08 formed significantly more and smaller nodules than did the wild type (Table 3). The acetylene-reduction activities (ARAs) of the nodules were measured to evaluate the nitrogen-fixing ability of the mutants. 5M09 and 5M14 had ARAs similar to that of the wild-type USDA110, but 5M08 and 20M19 had no ARA or much weaker ARA, respectively, than the wild-type USDA110 (Table 3). Large numbers of tiny nodules like those produced by 5M08 are often produced by Fix⁻ endosymbionts (16). We therefore considered that 5M08 and 20M19 had been subjected to mutations relevant to symbiotic nitrogen fixation. The remaining two mutants, 5M09 and 5M14, had significantly higher N₂OR activities than the parent strain USDA110, but they had normal symbiotic nitrogen fixation activity.

DISCUSSION

The positions of site-specific mutagenesis of the B. japonicum dnaQ gene (Fig. 1) were identical to those of the proofreading-deficient dnaQ926 mutant in the exonuclease I motif of the dnaQ gene of E. coli (9). When the mutated dnaQ was supplied in trans to B. japonicum USDA110 by pKQ2, the elevation of the mutation rate depended on the presence of pKQ2, as evaluated from the frequency of occurrence of spontaneous Km-resistant colonies (Tables 2 and 3) and the slower growth in the first cultivation cycle (Fig. 3A, left, and 4A and B, left). These phenotypes were similar to that of the E. coli dnaQ926 mutant (9, 22). Thus, it is likely that the dnaQ gene homolog (blr0640) in USDA110 encodes functional proofreading exonuclease by the epsilon subunit of DNA polymerase in B. japonicum.

By transconjugation and plant passage, we manipulated the introduction and elimination of plasmid pKQ2 with a modified dnaQ gene of B. japonicum USDA110 (Fig. 2). Cell populations showing increased salt stress or N₂OR activity were obtained exclusively from USDA110(pKQ2), indicating the occurrence of adaptive evolution of the mutator populations under the respective selection pressures. This may be due to the simultaneous occurrence of mutagenesis and selection during cell growth (37).

N₂O respiration selection pressure was different from that of salt stress. In the salt-stress experiment, mutant cells were selectively enriched after the seventh cultivation under high-salt conditions (100 mM NaCl) (Fig. 2 and 3B). In contrast, the enrichment processes in N₂O respiration were based on the assumption that cells with higher N₂OR activity grow faster than those with normal N₂OR activity. If so, the population of the former cells will dominate with time during selection. This was true for enrichment in an atmosphere of 5% (vol/vol) N₂O (Fig. 4A and Table 3). In contrast, 20% N₂O selection resulted in no significant differences in growth at the final cultivation cycle (Fig. 4). When we adopted selection in a 100% N₂O atmosphere, no growth was observed, even for wild-type USDA110 (data not shown). Thus, a low concentration of N₂O (5% vol/vol) is empirically important for selection based on N₂O respiration.

Among the four mutants with increased N₂OR activity, two (5M08 and 20M19) had weak or null fix phenotypes, and the remaining two (5M09 and 5M14) had normal levels of nitrogen fixation that were similar to that of wild-type USDA110 (Table 3). This suggested that mutator cells markedly accumulate unrelated mutations on their genomes.

There are several possible explanations for why the N₂OR activities of the four mutants were higher than that of the wild-type USDA110. The first possibility is that the expression level of the nos operon is upregulated by mutations in its promoter or related regulatory networks (21, 39), as has been observed after nos plasmid introduction (31). The second possibility is that a mutation in the nosZ gene produced an “improved” NosZ protein with increased N₂OR activity. A third explanation is the occurrence of a mutation in the electron carriers relevant to N₂OR activity (12). Because the mutants may accumulate many mutations on their genomes, as described above, simple DNA sequencing could not answer these questions. Genetic and biochemical studies are required to address these questions in relation to the respective mutants.

Soybean fields are sources of the emission of N₂O (18, 40). 5M09 and 5M14 had approximately 10 times the N₂OR activity of the wild-type USDA110 (Fig. 4D and Table 3). Because 5M09 and 5M14 no longer contain pKQ2 and the mutated dnaQ, they will not be considered genetically modified organisms. Thus, they could be released into agricultural fields to mitigate global warming (1). By using the pKQ2 system, it would be possible to reprepare B. japonicum mutants from indigenous isolates from local field soils. N₂OR activity and denitrification are environmentally important steps, and the mutated dnaQ method may be applicable to other environmental bacteria (3, 17).