Microbial populations in various paddy soils respond differently to denitrification-inducing conditions, albeit background bacterial populations are similar

Abstract

Rice paddy soil has been shown to have strong denitrifying activity. However, the microbial populations responsible for the denitrification in various rice paddy soils are not well known. In the present study, key denitrifiers were identified in six rice paddy soils by comparing the bacterial community structures in soils with strong denitrifying activity with those in soils without the activity. The effect of crop rotation on the denitrifier populations was also examined. Our study revealed three notable findings: (1) community structures of total bacteria were similar among the six rice paddy soils with different soil characteristics, (2) the populations that responded to denitrification-inducing conditions differed with soil type (Andosol, Gley soil or Gray lowland soil), (3) the populations that responded to denitrification-inducing conditions were similar within each type of soil, regardless of the cropping system (e.g. continuous or rotation).

Key words:

• community structure
• denaturing gradient gel electrophoresis
• denitrification
• rice paddy soil

Introduction

Denitrification is a microbial respiratory process in which nitrogen oxides (NO₃ ⁻ and NO₂ ⁻) are reduced stepwise to gaseous end products (NO, N₂O and N₂) (Philippot et al. 2007). Nitrous oxide (N₂O) is considered to be one of the major greenhouse gases (Philippot et al. 2007) and a most significant threat to ozone layer destruction (Ravishankara et al. 2009). Nitrous oxide is frequently produced as an end product of denitrification in upland crop fields; whereas little N₂O is emitted from rice paddy soils (Akiyama et al. 2006). The latter most likely occurs because of complete nitrate/nitrite reduction to N₂, as rice paddy soil is known to have strong denitrifying activity (Nishimura et al. 2004).

Recently in Japan, large areas of rice paddy fields have been restructured for efficient land and water management. This has been done in response to decreases in the population of farmers and in rice consumption in Japan. As restructured rice paddy fields drain water more efficiently, upland crops such as soybean can be grown on drained rice paddy fields. To adjust for the overproduction of rice, rotational cropping systems (e.g. 2 years of rice cropping followed by 2 years of soybean cropping) have been introduced to the restructured rice paddy fields (Nishimura et al. 2005). As a result, 84% of the soybean cultivation area in Japan was drained rice paddy soils in 2002 (Nishimura et al. 2005). Although emission of N₂O has been observed on upland crops grown on drained rice paddy fields (Nishimura et al. 2005), it is not clear if the denitrifier populations have changed with the change in the cropping system.

Various bacteria, as well as some archaea and fungi, have denitrifying ability (Hayatsu et al. 2008). Previous culture-independent studies have revealed that bacteria belonging to Burkholderiales and Rhodocyclales might be involved in denitrification in rice paddy soil collected from Nishitokyo City, Tokyo (Ishii et al. 2009b; Saito et al. 2008; Yoshida et al. 2009). However, it is unclear if these bacteria are also involved in denitrification in rice paddy soils collected from other areas because bacterial community structures might differ with location (Suzuki et al. 2009). Ishii et al. (2009b) showed that key denitrifiers can be identified by comparing community structures in soils with strong denitrifying activity with those in soils without denitrifying activity. The polymerase chain reaction-denaturing gradient gel electrophoresis (PCR-DGGE) technique has been shown to be useful in detecting populations responsive to specific conditions (Ishii et al. 2009a,b; Morimoto et al. 2008).

Consequently, the objectives of the present study were: (1) to characterize the bacterial communities of six rice paddy soils using PCR-DGGE analysis, (2) to identify key denitrifiers in these rice paddy soils by comparative analysis, (3) to examine the impact of a rotational cropping system on denitrifier populations.

Materials and methods

Soil samples

Sampling sites were located at the Kumamoto Prefectural Agricultural Research Center (32º89′N, 130º77′E), the Niigata Agricultural Research Institute (37º20′N, 138º45′E) and the Yamagata General Agricultural Research Center (38º25N, 140º24′E). At each site, two soil samples were collected from the top 10 cm of the rice field with different cropping systems: continuous rice cropping and rice–soybean rotation systems. The latter started in 2006 as a 4-year crop rotation system, including 2 years each of rice and soybean cultivations. Soybean and rice were grown in 2006–2007 and 2008–2009, respectively. Until 2005, rice had been grown in these fields. Sampling was done in late April to early May (before water irrigation) in 2008. Soil samples were sieved through 2 mm mesh and stored at 4°C in the dark until required (∼1 month). The physicochemical characteristics of the soil samples are shown in Table 1.

Table 1 Physicochemical characteristics of the soil samples

Sampling location	Cropping system	Soil type	Soil texture	pH (H2O)	Total C (g kg^−1)	Total N (g kg^−1)
Kumamoto	Continuous rice	A	SiL	6.8	34.5	2.7
Kumamoto	Rice–soybean rotation	A	SiL	7.0	36.0	8.1
Niigata	Continuous rice	GS	CL	5.6	14.0	1.27
Niigata	Rice–soybean rotation	GS	CL	6.0	11.9	1.15
Yamagata	Continuous rice	GLS	CL	5.7	21.9	1.7
Yamagata	Rice–soybean rotation	GLS	CL	5.9	22.4	1.7
A, Andosol; CL, Clay loam; GLS, Gray Lowland Soil; GS, Gley Soil; SiL, Silt loam.

Soil microcosm

A previously established reproducible soil microcosm setup (Ishii et al. 2009b; Saito et al. 2008) was used in the present study. In brief, 1 g of fresh soil was placed in a 10-mL glass serum vial (Nichiden-Rika Glass, Kobe, Japan) and submerged in sterile water for 1 week at 30°C (pre-incubation) to simulate the water irrigation process. After pre-incubation, 1.8 mL of the upper soil-free water was removed and 0.1 mg N-nitrate and/or 0.5 mg C-succinate was added. The vial was anaerobically incubated with Ar–C₂H₂ (90:10) gas at 30°C for 24 h. For each soil, three treatments were examined. In the NS treatment the vials were incubated with added nitrate and succinate, whereas in the SU treatment the vials were incubated with only added succinate. In the BA treatment the vials were not supplemented with any substrate and were not incubated. Three replicate vials were prepared for each treatment. The production of CO₂ and N₂O was measured after 24-h incubation in each vial by using gas chromatography as described previously (Saito et al. 2008). Incubation with nitrate and succinate has been shown to greatly enhance denitrification activity (Saito et al. 2008). Previous studies have also shown that succinate can be used by various denitrifiers, but is rarely used for other functions, such as fermentation, metal reductions and dissimilatory nitrate reduction to ammonium (DNRA), within 24-h incubation in the presence of nitrate (Ishii et al. 2009b; Saito et al. 2008). In addition to denitrification, nitrate respiration, which converts NO₃ ⁻ to NO₂ ⁻ and does not produce gaseous end products, may also occur under the conditions used (Ishii et al. 2009b).

DNA extraction, PCR-DGGE and sequencing analyses

The DNA was extracted from the soil (0.8 g) by using an ISOIL for Beads Beating kit (Nippon Gene, Tokyo, Japan) as previously described (Ishii et al. 2009b). The PCR was carried out with 357F-GC and 520R primers (Ishii et al. 2009b; Muyzer et al. 1993) to amplify the V3 region of the 16S rRNA gene. Aliquots of 750 ng of the PCR products were loaded onto 8% acrylamide gel with 35–65% denaturant concentrations and electrophoresis was carried out at 60 V for 800 min. The gel was stained with SYBR Green II (Invitrogen, Carlsbad, CA, USA) and the gel image was captured using a Printgraph GX CCD camera (ATTO, Tokyo, Japan). Principal component analysis (PCA) was carried out based on the digitized DGGE banding profiles by using R software version 2.9.0. (http://www.r-project.org) as previously described (Ishii et al. 2009a).

The DGGE bands that were unique to the samples amended with nitrate and succinate were excised from the gel by using a razor blade. The DNA was eluted from the gel and used for PCR with 357F and 520R primers (Muyzer et al. 1993; Otsuka et al. 2008). The PCR products were ligated into the pGEM-T Easy vector system (Promega, Madison, WI, USA) and transformed into Escherichia coli JM101 competent cells according to the manufacturer’s instructions. Cloned insert DNA were amplified by PCR with vector primers T7-1 and SP6 and used for the sequencing reaction as described previously (Saito et al. 2008). Four clones per DGGE band were sequenced and their taxonomic assignment was examined using Naïve Bayesian Classifier version 2.0 (Wang et al. 2007). Nucleotide sequences of multiple clones were aligned and a neighbor-joining tree was constructed using the ClustalW program as previously described (Ishii et al. 2009b). The nucleotide sequences determined in the present study were deposited into DNA Data Bank of Japan under accession numbers AB535663–AB535679.

Results and discussion

The DGGE banding profiles were similar within the replicates (data not shown), as found in a previous study (Ishii et al. 2009b). Therefore, we combined each set of replicate DNA samples to compare the bacterial community structures of the six soil samples with the three treatments (total of 18 DNA samples) in a single denaturing gradient gel. The banding profiles showed that the bacterial community structures were similar among the six soils in the BA and SU treatments (Fig. 1a). These samples were also tightly clustered in a PCA plot (Fig. 1b), suggesting that the bacterial community structures were similar under paddy conditions, regardless of soil type. Similar results were obtained from several field rice paddy soils using DGGE analysis (Hiramatsu et al. 2007).

Samples amended with nitrate and succinate (NS treatment) had different community structures than the BA and SU treatments (Fig. 1), suggesting that several groups of bacteria may have been enriched by the NS treatment. Denitrification activity, as measured by the accumulation of N₂O gas under the presence of 10% C₂H₂, was detected in these NS-treated samples (data not shown). When vials were incubated with 100% Ar gas, instead of Ar–C₂H₂ (90:10) gas, no N₂O accumulated (data not shown), suggesting that denitrifiers in the NS-treated samples reduced NO₃ ⁻ completely to N₂ gas. These results indicate that the bacteria represented by the distinct DGGE bands (shown by arrows in Fig. 1a) may have increased their population by performing complete denitrification. Interestingly, these enriched populations in the NS-treated samples differed between the soils, although the background bacterial community structures were similar. Based on the sequence analysis of the excised bands, clones related to the genus Chromobacterium were common in the NS-treated Kumamoto and Niigata soils; whereas clones related to the genus Ammoniphilus were common in the NS-treated Niigata and Yamagata soils (Table 2). Clones obtained from different soils were phylogenetically closely related, and some had 100% identical sequences (Fig. 2). In contrast to Chromobacterium and Ammoniphilus, clones related to the order Rhodocyclales (e.g. Dechloromonas sp. and Azoarcus sp.) were only seen in Kumamoto soil. Previous studies have shown that members of Burkholderiales and Rhodocyclales respond to the addition of nitrate and succinate in rice paddy soil collected from Tanashi, Tokyo (Ishii et al. 2009b; Saito et al. 2008). As both the Tanashi and Kumamoto soils were classified as Andosol, the appearance of Rhodocyclales bacteria in response to additions of nitrate and succinate may be related to the characteristics of the Andosol. Similar to our study, Suzuki et al. (2009) reported that bacterial communities in soil grown with upland crops were more dependent on soil type (e.g. Andosol or grey lowland soil) than fertilizer type (e.g. chemical fertilizer or manure). As Andosol is derived from volcanic ash, this soil is well-drained compared with soils of alluvial origin (e.g. Gley soil and Gray-lowland soil) (Nishimura et al. 2005). In addition, Andosol is usually rich in organic matter, as is the case with the Kumamoto soil. These Andosol characteristics may have selected Rhodocyclales, which is a facultative–anaerobic and metabolically versatile bacteria (some Rhodocyclales can degrade recalcitrant organic matter) (Loy et al. 2005), as major denitrifiers responsive to the NS treatment. More Andosol samples need to be examined to draw further conclusions about the appearance of Rhodocyclales bacteria. Collectively, these results indicate that Chromobacterium sp. and Ammoniphilus sp. are common denitrifiers in several rice paddy soils, whereas some other denitrifiers (e.g. Dechloromonas sp. and Azoarcus sp.) are more specific to a particular soil, under the conditions used in the present study. As 16S rRNA gene-based analysis alone cannot identify the ecological role of bacteria, isolation of these bacteria is necessary to confirm that they perform denitrification for growth.

Figure 1 Bacterial community structures assessed by denaturing gradient gel electrophoresis (DGGE) analysis. (a) The DGGE banding profiles of the six soil samples and three treatments. Lanes 1–3, soil sample collected from the continuous rice field from Niigata (NC); lanes 4–6, soil sample collected from the rice–soybean rotation field from Niigata (NR); lanes 7–9, soil sample collected from the continuous rice field from Yamagata (YC); lanes 10–12, soil sample collected from the rice–soybean rotation field from Yamagata (YR); lanes 13–15, soil sample collected from the continuous rice field from Kumamoto (KC); lanes 16–18, soil sample collected from the rice–soybean rotation field from Kumamoto (KR); BA, samples without incubation; SU, samples incubated with added succinate; NS, samples incubated with added nitrate and succinate. (b) Principal component analysis (PCA) plot based on the DGGE profile. The normalized location and intensity of each DGGE band were used in the PCA analysis. The numbers in the plot correspond to the lane numbers in (a). The numbers in parentheses are the percentages of variation explained by each component.

Table 2 Taxonomic assignment of the clones obtained from the excised denaturing gradient gel electrophoresis bands

Sample	DGGE band	Genus (Order)
NC	a	Ammoniphilus (Bacillales), Chromobacterium (Neisseriales)
	b	Ammoniphilus (Bacillales)
	c	Chromobacterium (Neisseriales)
NR	d	Bradyrhizobium (Rhizobiales), Chromobacterium (Neisseriales)
	e	Chromobacterium (Neisseriales)
	f	Chromobacterium (Neisseriales)
YC	g	Ammoniphilus (Bacillales)
	h	Ammoniphilus (Bacillales)
YR	i	Ammoniphilus (Bacillales)
	j	Ammoniphilus (Bacillales)
KC	k	Dechloromonas (Rhodocyclales)
	l	Chromobacterium (Neisseriales)
KR	m	Azoarcus (Rhodocyclales), Dechloromonas (Rhodocyclales)
	n	Chromobacterium (Neisseriales)
DGGE, denaturing gradient gel electrophoresis.

Figure 2 Neighbor-joining tree showing the phylogenetic relationship between the clones obtained from the excised denaturing gradient gel electrophoresis band shown in Fig. 1. Bootstrap values (%) were generated from 1,000 replicates and only values >70% are shown. α, Alphaproteobacteria; β, Betaproteobacteria; B, Bacilli.

Our results also suggest that the cropping system (e.g. continuous or rotation) did not substantially influence the denitrifier populations. Similar populations responded to the NS treatment in the continuous and rotational cropping systems (Figs 1,2), indicating that the same groups of bacteria may carry out denitrification under the conditions used in the present study. This result implies that the denitrifiers acting under paddy conditions may have survived in the following 2 years of soybean cultivation (rotational cropping) and re-emerged when conditions become favorable. Although denitrification and the resulting N₂O emission could occur during soybean cultivation in fields converted from rice paddy (Nishimura et al. 2005), it is not clear from our study if N₂-generating denitrifiers under rice paddy conditions can produce N₂O under well-drained conditions. This needs to be clarified in a future field study.

In conclusion, our results suggest that the fundamental bacterial populations were similar among the rice paddy soils examined in the present study, but that the denitrifier populations might be affected by soil type. In addition, our results suggest that the impact of the cropping system on denitrifier populations is minimal. These results are important when establishing strategies to mitigate N₂O emission from upland crop fields converted from rice paddy. As our results were obtained using a soil microcosm experiment, further study is necessary to investigate whether the “key denitrifiers” identified in the present study really perform denitrification in field rice paddy soils by using rRNA-based and/or mRNA-based analysis.

Acknowledgments

We thank Shigeto Sudo and Kazuyuki Yagi (National Institute for Agro-Environmental Sciences), Hiroyuki Mizukami and Noriaki Gunjikake (Kumamoto Prefectural Agricultural Research Center), Yuichiro Furukawa and Yutaka Shiratori (Niigata Agricultural Research Institute) and Masamichi Adachi, Yuichiro Hurukawa, Hiroshi Saito, Hiroyuki Shiono and Katsuki Kumagai (Yamagata General Agricultural Research Center) for their help in collecting soil samples. This work was supported by the Program for the Promotion of Basic Research Activities for Innovative Biosciences (PROBRAIN) from the Bio-oriented Technology Research Advancement Institution, Tokyo, Japan. Additional financial support was provided from a Grand-in-Aid (Soil eDNA) from the Ministry of Agricultural, Forestry and Fisheries of Japan.