Introduction

The interaction between land-use changes and greenhouse gas emission has been studied because of its importance to develop future scenarios of climate and biodiversity (Lonborg et al. 2009; Wu et al. 2009; Attard et al. 2011). Among these emissions, N2O emission from soils results from nitrification and denitrification (Cheng et al. 2009). Understanding the controlling factors of these emissions is crucial for quantifying the effects of land-use change on the N cycle in terrestrial ecosystems and for managing and mitigating the environmental consequences associated with N pollution (Boyer et al. 2006). Numerous studies have been conducted on the denitrification process in soil, especially its role in N losses and how it is affected by environmental factors (Galloway et al. 2004; Hofstra and Bouwman 2005; Boyer et al. 2006; Attard et al. 2011). However, denitrification is measured mainly in soils from temperate zones in industrialized countries (Hofstra and Bouwman 2005) and few measurements are available on denitrification rates in soils from subtropical regions (Xu and Cai 2007; Zhang et al. 2009).

Many humid subtropical soils are highly acidic and strongly weathered (Bennema et al. 1970), and in humid subtropical forest soils, nitrification and denitrification rates are low (Zhao et al. 2007). However, subtropical soils are considered to be the major natural sources of NO and N2O, which are both produced by nitrification and denitrification (Lashof and Ahuja 1990; Matson and Vitousek 1990; Davidson 1991; Davidson and Kingerlee 1997; Stehfest and Bouwman 2006). Both soil organic C and soil N availability are the key factors affecting denitrification in subtropical soils of China (Xu and Cai 2007), but this was determined by changes in nitrate content of red soils, which is an inaccurate method. Recently, potential denitrification rate has been determined by changes in total N gas rate (Zhang et al. 2009; Attard et al. 2011; Sun et al. 2012).

Denitrification consists of four reactions catalyzed by nitrate reductases, nitrite reductases, nitric oxide reductases, and nitrous oxide reductase. The nitrite reductase is the key enzyme of denitrification, since it catalyzes the reduction of nitrite to gas products (Henry et al. 2004), that is the reaction limiting the rate of the denitrification process (Zumft 1997). Several studies have used narG and/or napA, nirK and/or nirS, norB, and nosZ as molecular markers of the denitrifying bacteria to study their diversity in various environments (Gregory et al. 2000; Henry et al. 2006; Kandeler et al. 2006; Smith et al. 2007). Both NO and N2O are the intermediates of the denitrification pathway, and their emission rates depend on the balance of their production and consumption. For example, it has been reported that Nos synthesis and activity can be repressed and inhibited to a greater extent by O2 compared to the other reductases (Otte et al. 1996). If N2O production exceeds N2O consumption, N2O can diffuse away from the site of activity (Firestone et al. 1980). Therefore, increases in the N2O/(NO+N2O+N2) ratio reflect greater rate of NO reduction to N2O (driven by norB) than that of N2O reduction to N2 (driven by nosZ). Generally, these studies have been conducted in temperate soils (Boyle et al. 2006; Katsuyama et al. 2008), whereas the relationship among denitrifier abundance, soil potential denitrification, and total N gas production in soils of subtropical regions as affected by the different land use is poorly known. Therefore, the aims of our research are to study the following: (1) the responses of potential denitrification and total N gas production to land-use changes in subtropical soils, and (2) the key soil variables affecting potential denitrification and total N gas production in different land-use soils of the subtropical region. For these reasons, we sampled soils of conifer forest, shrub forest, and farmland derived from two typical parent materials of granite and Quaternary red earth in the subtropical China. Sampled soils were analyzed for their soil physicochemical properties, potential denitrification, NO, N2O, and N2 production, and abundance of denitrification genes (i.e., narG, nirK, norB, nosZ).

Materials and methods

Soil sampling

Sampling sites were located in Yingtan City (27°52′∼28°17′ N, 116°56′∼117°30′ E), Jiangxi province, a typical subtropical region of southern China. The mean annual temperature and precipitation are 18.4 °C and 1,785 mm, respectively. The soils are acrisols and ferralsols (Doran and Zeiss 2000), which are predominantly derived from Quaternary red earth (6.48 %), tertiary red sandstone (20.38 %), and granite (21.92 %) (Zhao et al. 1988). The dominant clay minerals of soils are kaolinite and hydrous mica with presence of vermiculite. The natural vegetation consists of a mixed forest of oak (Quercus mongolica), Chinese red pine (Pinus massoniana Lamb), and China fir (Cunninghamia lanceolata (Lamb.) Hook.) (Xu et al. 2003), and now the land is mainly covered by forest trees, shrub, and all kinds of crops (e.g., peanut, rice paddy, etc.), which account for 51.4, 1.2, and 12.8 % of the land, respectively (Editorial Committee of ECA 2005). Chemical fertilizers, such as urea and diammonium phosphate, are intensively applied to farmland (Zhu and Chen 2002). In our study, we have considered three land uses, the natural secondary coniferous forest (C), shrub forest land (S), and farmland (F) samples, which were collected from the Quaternary red earth (Q) and granite (G) area. Totally, six soils, GC, GS, GF, QC, QS, and QF, were sampled from sites with similar slope. In each soil, three random plots (each 100 × 100 m) were selected with a minimum distance of 100 m. From each plot, three subsamples were taken from the A horizon (0–20 cm) and then mixed to obtain one composite sample. Therefore, totally 18 soil samples were collected on October 2011. Fresh soil was sieved through a 2-mm mesh with removal of visible stones and vegetable. One subsample was air-dried for chemical analyses. One fresh subsample was stored at 4 °C before analysis, whereas the residual subsample was immediately stored at −80 °C for molecular analysis.

Anaerobic incubation

The denitrification rate was determined by the anaerobic incubation method of Zhang et al. (2009), with the following modifications: (1) acetylene (C2H2) was used to inhibit N2O reduction to N2; (2) each time gas fluxes were measured, the flasks were flushed with oxygen-free N2 gas, resealed with stoppers and incubated for 3 h; and (3) soil NO3 −−N concentration was adjusted to 50 mg NO3 −−N kg−1 dry soil weight in each flask.

For each soil sample, a series of 250-ml Erlenmeyer flasks was prepared with 10 g of soil (oven-dry basis) and 10 ml deionized water. To achieve 50 mg NO3 −−N kg−1 dry soil weight, a suitable amount of potassium nitrate solution containing 10 mg NO3 −−N was added to each flask, which was immediately capped with an alright silicone rubber stopper fitted with butyl rubber septum. Silicone sealant was used around the stopper to ensure anaerobic conditions. The flasks were connected to a multiport vacuum manifold, which allowed 20 flasks to be simultaneously vacuumed and flushed with oxygen-free N2 gas. The procedure was repeated three times (each for 10 min) to create an anaerobic headspace. Then 25 mL C2H2 were used to immediately replace an equal amount of headspace air for five flasks of each treatment and pumped several times for an even distribution in the substrate, water and headspace, whereas the other five flasks of each treatment contained no C2H2. After sealing with butyl rubber, all flasks were shaken for 60 min on a 250-rpm mechanical shaker at 25 °C. Following equilibration at atmospheric pressure, the flasks were incubated at 25 °C in the dark. The N2O and NO concentrations of each soil sample were measured by randomly selecting five flasks with C2H2 addition and five flasks without C2H2 addition and sampling 25 ml headspace gas at 3, 27, 51, 75, and 99 h.

Analyses of soil physicochemical properties and nitrogen gasses

Soil pH was measured in a soil to water ratio of 1:5 (v/v) using a DMP-2-mV/pH detector (Quark Ltd, Nanjing, China). Soil organic C (SOC) was determined by wet digestion with H2SO4–K2Cr2O7 (Bremner and Jenkinson 1960). Total N was determined by semimicro Kjeldahl digestion using Se, CuSO4, and K2SO4 as catalysts (Bremner 1960). Available N was determined by alkali-hydrolyzed diffusing method (Bronner and Bachler 1980). Total P was determined by the alkali-melting molybdenum blue method (Bray and Kurtz 1945). Total K was determined by alkali-melting atomic absorption spectrometry, and available K was extracted by ammonium acetate and determined by atomic absorption spectrometry (Zagatto et al. 1979). Inorganic N (NO3 −, NO2 −, and NH4 +) was extracted with a 2 M KCL solution at a soil/water ratio of 1:5; the solution was shaken (250 rpm) for 60 min at 25 °C and NO3 −, NO2 −, and NH4 + contents were determined by a continuous flow analyzer (Skalar, Breda, the Netherlands).

The N2O concentration was determined using a 2-mm ID stainless steel column, 3-m long, packed with Porapak Q (80/100 mesh), and an Agilent 7890 gas chromatograph fitted with an electron capture detector set at 300 °C (Zhang et al. 2011; Zhu et al. 2012). The column temperature was maintained at 40 °C, and the carrier gas was argon–methane (5 %) at a flow rate of 30 ml/min. The N2 emission was obtained by subtracting the N2O emission without acetylene from the N2O emission with acetylene (Ryden et al. 1979; Zhu et al. 2012). For the determination of NO, a gas sample drawn from a flask was diluted to 1,000 ml using highly purified N2 and the sample concentration measured using an EC 9841 T NOx analyzer (ECOTECH, chemiluminescence detector).

Soil DNA extraction and quantitative polymerase chain reaction

DNA was extracted from about 0.5 g fresh soil using the Fast DNA SPIN kit for soil (MP Biomedicals) following the manufacturer’s instruction. The quality and quantity of DNA were checked using a NanoDrop spectrophotometer (NanoDrop Technologies Inc., Wilmington, DE). Amplification of quantitative PCR products was carried out with a CFX96 Optical Real-Time Detection System (Bio-Rad Laboratories, Inc., Hercules, CA) by using SYBR green as the detection system in a reaction mixture of 25 μL containing 0.5 μM (each) primer; 12.5 μL of SYBR green PCR master mix, including HotStar Taq DNA polymerase, QuantiTect SYBR green PCR buffer, deoxynucleoside triphosphate mix with dUTP, SYBR green I, ROX, and 5 mM MgCl2 (QuantiTect SYBR green PCR Kit; QIAGEN, France); and 1.25 μL of DNA-diluted template corresponding to 12.5 ng of total DNA. Bovine serum albumin (500 ng/reaction; Takara, Dalian, China) was used to enhance PCR efficiency. Thermal cycling conditions for the narG, nirK, norB, and nosZ genes were those previously described (Henry et al. 2004, 2006). The nirK quantitative PCR was performed with the nirKF1aCu (ATCATGGTSCTGCCGCG) and nirKR3Cu (GCCTCGATCAGRTTGTGGTT) primers. The used nosZ primers were nosZ2F (CGCRACGGCAASAAGGTSMSSGT) and nosZ2R (CAKRTGCAKSGCRTGGCAGAA). The narG primers were narG1960F (TAYGTSGGSCARGARAA) and narG2650R (TTYTCRTACCABGTBGC), and norB amplification was performed with the cnorB BF (AIGTGGTCGAGAAGTGGCTCTA) and cnorB BR (TCTGIACGGTGAAGATCACC) primers. Three independent Q-PCRs were performed for each gene and each soil replicate. Standard curves were obtained with serial plasmid dilutions of a known amount of plasmid DNA containing a fragment of the narG, nirK, norB, or nosZ gene (Henry et al. 2004, 2006). Soil DNAs were also tested for the inhibitory effects of coextracted substances by determining the functional gene copy number in tenfold dilutions of soil DNA. In addition, standard plasmid DNA was quantified by Q-PCR with and without the addition of environmental DNA, and the obtained values were compared.

Calculation of denitrification rate

The denitrification rate was expressed as the total N gas production (NO+N2O+ N2) per hour during whole incubation periods (Zhang et al. 2009). The cumulative total N gas production linearly increased with time in the initial 99 h of incubation in all studied soils (average R 2 was 0.97), indicating total N gas production was not limited by nitrate concentration or by other factors affecting denitrification; the slope of cumulative total N gas production with time indicated the potential denitrification rate.

Statistical analysis

Pearson’s correlation analysis was applied to correlate edaphic variables with denitrification rate by using the SPSS software package 16.0 for Windows. Differences in soil properties and denitrification rate among different land-use soils were assessed by ANOVA analysis and compared with a Tukey’s test at p = 0.05.

Multiple stepwise regression analyses were applied to determine the soil variables (i.e., soil organic C, total N, total P, total K, available N, available K, the C/N ratio, pH, the initial NH4 +−N, and NO3 −−N content) significantly related to the potential denitrification rate. Stepwise criteria involved the use probability of F to enter at p ≤ 0.05 and to remove at p ≥ 0.1. The confidence intervals were 95 %. The standardized coefficients, R square and adjusted R square were used to judge the importance of each variable to the potential denitrification rate. Differences in the edaphic factors and denitrification rates among soils were assessed by ANOVA analysis and compared with a Tukey’s test at p = 0.05.

Results

Soil properties

Soil physicochemical properties differed significantly under different land-use soils (Table 1). Soil pH values ranged from 4.41 to 4.67 in forest soils (GC, QC, GS, QS) and were lower than in farmland soils (GF, QF), varying from 5.02 to 5.28. Soil organic C concentration and C/N ratio of forest soils (GC, QC, GS, QS) were higher than in farmland soils (GF, QF). Total N concentration in forest soils of granite (GC, GS) was higher than in farmland soils (GF), while there was no difference among QC, QS, and QF soils. Soil total P concentrations were significantly lower in forest soils than in farmland soils, mainly due to artificial fertilization. There were no significant differences for both NH4 +−N and NO3 −−N contents among different land-use soils.

Table 1 Selected physicochemical properties of the soils (0–20 cm) developed from Quaternary red earth or from granite and subjected to different land uses of the following: conifer forest, shrub forest, and farmland
Full size table

Change in inorganic N during incubation

The changes in the NH4 +−N content were different between forest and farmland soils of the same parent material (Fig. 1a). The increment in NH4 +−N concentration from the beginning to the end of incubation in GC, GS, GF ,QC, QS, and QF were 6.84, 6.28, 3.97, 6.46, 5.39, and 7.93 mg kg−1, respectively. The content of NO2 −−N of all studied soils was low (<2.5 mg kg−1) and changed throughout anaerobic incubation (Fig. 1b). The content of added NO3 −−N was reduced with increased incubation time under anaerobic incubation, but the reduction rates differed among different land-use soils (Fig. 1c). The decrease in NO3 −−N concentration from the beginning to the end of incubation in GC, GS, GF ,QC, QS, and QF were 14.11, 18.61, 31.12, 7.38, 3.58, and 14.59 mg kg−1, respectively.

Fig. 1
figure 1

Changes in NH4 +−N (a), NO2 −−N (b), and NO3 −−N (c) content during the anaerobic incubation. GC, GS, and GF represent red soils of coniferous Pinus forest, bush forest, and farmland (rice paddy) with granite as parent material, respectively; QC, QS, and QF represent red soils of coniferous Pinus forest, bush forest, and farmland (peanut) with Quaternary red earth as parent material, respectively. Bars are standard deviation values

Full size image

Nitrogen gas emission rates and potential denitrification rate

During anaerobic incubation, the N gas emission rates varied and differed among the six soil samples (Fig. 2); N2, NO, and N2O emission of forest soils of granite (GC, GS) were lower than in farmland soils (GF). Total denitrification gas (NO, N2, and N2O) increased linearly with incubation time in the studied soils (R 2 values were 0.95, 0.98, 0.99, 1.00, 1.00, and 0.98 for GC, GS, GF, QC, QS, and QF soils, respectively).

Fig. 2
figure 2

NO (a), N2 (b), and N2O (c) emission rate in the soil samples during anaerobic incubation. GC, GS, and GF represent red soils of coniferous Pinus forest, bush forest, and farmland (rice paddy) with granite as parent material; QC, QS, and QF represent red soils of coniferous Pinus forest, bush forest, and farmland (peanut) with Quaternary red earth as parent material, respectively. Bars are standard deviation values

Full size image

The average denitrification rates in GC, GS, GF, QC, QS, and QF were 0.046, 0.053, 0.122, 0.008, 0.008, and 0.021 mg kg−1 h−1, respectively. Therefore, for the same parent material soil, the potential denitrification rates in farmland soils were higher than those of forest soils (p < 0.05); the rates in granite soils were higher than those in Quaternary red earth soil under the same land use pattern (p < 0.05). There was a significant relationship between potential denitrification rate and soil pH (y = 0.063x − 0.251, R 2 = 0.26, p < 0.05). But no significant relationships were found between denitrification rate and content of soil organic C and total N.

Soil denitrifying microorganism

Abundance of narG gene of farmland soils was higher than in forest soils under the same soil parent material (p < 0.05) (Fig. 3a). However, the abundance of nirK gene did not significantly differed among difference land-use soils, except QF had more nirK gene copies than other soils (p < 0.05) (Fig. 3b). Both norB and nosZ genes showed a similar pattern being higher in farmland than in forest soils deriving from the same parent material (p < 0.05), and no significant difference was found between two parent material soils under the same land-use pattern (p > 0.05) (Fig. 3c, d). Soil pH was positively correlated with the abundance of norB gene (y = 1.28 x −5.54, R 2 = 0.34, p < 0.05) and nosZ gene (y = 5.08x − 21.98, R 2 = 0.54, p < 0.05). The abundance of norB gene was significantly correlated with the N2 emission rate (y = 0.042x + 0.007, R 2 = 0.80, p < 0.05) and denitrification rate (y = 0.006x + 0.021, R 2 = 0.73, p < 0.05).

Fig. 3
figure 3

narG (a), nirK (b), norB (c), and nosZ (d) gene copies in six soil samples. GC, GS, and GF represent red soils of coniferous Pinus forest, bush forest, and farmland (rice paddy) with granite as parent material, respectively; QC, QS, and QF represent red soils of coniferous Pinus forest, bush forest, and farmland (peanut) with Quaternary red earth as parent material, respectively. Bars are standard deviation values

Full size image

Discussion

Denitrification rate in different land-use soils

Denitrification rates of forest soils were lower than in farmland soils in the subtropical region of China. Many factors can affect the denitrification rate of soil, including soil water content, temperature, NO3 − concentration, available organic C, pH, redox potential, and microbial activity (Simek and Cooper 2002; Galloway et al. 2004; Heinen 2006). Since we surveyed soil properties (e.g., pH, SOC, total N, and C/N ratio) and determined the denitrification rate by using an anaerobic incubation method, the relationship among these parameters can be evaluated. Multiple regression analysis indicated that soil pH was the best predictor of the denitrification rate of the studied soils (R 2 = 0.23, p < 0.05, n = 18). Both neutral and alkaline pH values are optimal conditions for denitrification, while denitrification capacity is low in acidic soils (Wijler and Delwiche 1954). Indeed, the potential denitrification rates were significantly correlated with soil pH (p < 0.05). Also, Sun et al. (2012) have found that denitrification of acidic soils along Northeast China Transect depended on soil pH. However, contradictory results have been published on the relationship between soil pH and denitrification (Simek and Cooper 2002). Xu and Cai (2007) reported that SOC and total N were the controlling factors affecting denitrification in the subtropical soils of China, probably because some of tested soils derived from sandstone with a pH > 5.0 had no denitrification ability. However, Xu and Cai (2007) determined denitrification by the reduction of nitrate content, which is not an accurate method (Zhang et al. 2009), as already stated.

Denitrification can depend on the content of soil organic C (Heinen 2006; Jäger et al. 2011), but we did not find a significant relationship between SOC and denitrification rate in the studied soils. This may be due to the fact that the amount of SOC in the studied soils is sufficient to supply the amount of organic C needed for denitrification in the short-term (99 h) being low denitrification of these soils (Xu and Cai 2007; Zhang et al. 2009). The amount of soluble C showed high variability in controlling the denitrification rate than soil pH (Koskinen and Keeney 1982; Parkin et al. 1985; Fang et al. 2009). Unfortunately, the content of soluble C was changeful in the studied region (Zhou et al. 2002), so it was not determined and evaluated.

Microbial mechanism of denitrification

The synthesis and activity of denitrifying enzymes and the relative gas production are affected by soil properties, such as soil pH, soil organic C, oxygen partial pressure, and soil redox potential (Kralova 1991; Manconi et al. 2006; Chapuis-Lardy et al. 2007; Attard et al. 2010). The abundance of both norB and nosZ gene copies were positively correlated to soil pH. Therefore, soil pH also affected the presence of some genes codifying denitrifying enzymes, confirming what reported by Wolf and Brumme (2003) and Dannenmann et al. (2008).

It is well known that changes in potential denitrification depend on (1) changes in the physiological activity of individual cells (i.e., mean activity per denitrifying cell), (2) changes in denitrifier abundance, and (3) changes in the denitrifier community composition (Attard et al. 2011). Actually, denitrification is an inducible activity with key enzymes being synthesized according to environmental conditions (Zumft 1997). Therefore, potential denitrification and denitrifier abundance can be affected by soil properties and different land uses (Xu and Cai 2007; Attard et al. 2011). The genetic structure of the denitrifying community, as evaluated by targeting narG, nirK, norB, and nosZ genes was significantly affected by changes in land use and soil parent material. In addition, potential denitrification was positively correlated with the abundance of norB gene copies but not with the abundance of narG, nirK, and nosZ gene copies. Probably, the reduction of NO to N2O is the key denitrification process in these subtropical soils. In contrast, other studies reported that changes in denitrification rate did not depend on changes in the denitrifier community structure (Henry et al. 2008; Hallin et al. 2009; Attard et al. 2011). A further understanding of this topic requires detecting active populations in the denitrification by RNA-based approaches targeting active populations.

In conclusion, the potential denitrification and total N gas products differed significantly among soils sampled from different land uses. The soil pH and the abundance of the NO reductase gene (norB) seem to play an important role in the denitrification of the subtropical soils of China, but further research is needed to understand the underlying mechanisms.