Introduction

Factors driving microbial N cycle in soils during the vegetation period have been studied in detail in the last decades. It has been clearly shown that many factors such as mineral N and available C contents, oxygen concentration, soil water content, pH value, and temperature (Davidson and Kingerlee 1997) influence the N transformation rates. Although high denitrification rates during the winter months have been observed accounting for up to 30% of the annual N2O emissions from agricultural soils (Ruser et al. 1998), there is still a lack of understanding how N turnover is regulated during the cold time of the year (Phillips 2008).

Legumes and legume–grass mixtures, frequently used as green manure in organic farming, might be one factor influencing N turnover in winter. These plants are cropped in autumn, remain on the fields over winter, and supply plants in the following vegetation period with N. Due to the low C/N ratio of green manures compared to other plant residues (e.g., straw of cereals, which is left on the fields after harvest in autumn), more N is also available for microbial activities. However, it is poorly understood how and under which conditions the N of the plant residues is transformed during the winter months. We expected that due to freeze–thaw processes and the increased amounts of available C, e.g., microbes involved in denitrification will become more abundant depending on the quality and amount of applied N.

Therefore, we investigated the influence of two common agricultural plant residues (wheat straw and lucerne-clover-grass-mix) considerably differing in their C quality and C/N ratio on nitrifiers and denitrifiers based on abundance of the functional genes for the ammonia monooxygenase gene amoA and the nitrite reductase genes nirK and nirS, respectively, and on mineral N fractions (NO −3 and NH +4 ) after a freeze–thaw cycle.

Materials and methods

The experimental site is located on the research farm “Klostergut Scheyern” in Southern Germany (48°30.0′ N; 11°20.7′ N). Soil samples were taken from an arable field (Cambisol) in September 2008 several weeks after harvest of summer wheat. The upper 20 cm of the plowed soil horizon were sampled and characterized as follows: 22% clay, 36% silt, and 42% sand; 1.5% organic C, C/N ratio of 10.0; pH value 5.8 (0.01 M CaCl2). The soil was sieved at 2 mm and air-dried at room temperature.

Wheat straw and growth of a lucerne-clover-grass-mix were chosen as organic amendment reflecting common agricultural situations at field sites in autumn and winter. The different plant residues exhibited similar C contents (43.6% and 42.1%, respectively) but differed considerably in their C/N ratios (126.6 and 10.1, respectively), recalcitrance, and biodegradability (Mueller et al. 1998; Nicolardot et al. 2007; Wang et al. 2004). The residues were air-dried and ball-milled (Retsch MM2, Germany) before application.

The experiment was set up in the following way: 50 g soil was amended with 2 g of the respective plant residues and incubated in cylinders of 4-cm height and a volume of 100 cm3. Thus, similar amounts of C were added to the soil (corresponding to 222 dt C ha−1), however of different quality and C/N ratios. Per treatment and sampling time point, four independent replicate cylinders were prepared. All samples were equilibrated during a pre-incubation period of 2 weeks at 10°C and 60% water holding capacity. Half of the samples were then frozen for 3 days at −20°C and afterwards returned to 10°C; the other half of the samples remained constant at 10°C. Sampling was carried out at day 0 (day before freezing) and days 4, 6, and 10 (1, 3, and 7 days after thawing), respectively.

Aliquots of soil samples were immediately extracted with 0.01 M CaCl2 [soil/CaCl2 ratio 1:2 (wt/wt)] according to Zsolnay (2003). Both NH +4 –N and NO −3 –N concentrations were measured using the Nanocolor Ammonium 3 and Nanocolor Nitrat 50 kit (Merck, Germany), respectively. The concentrations of water-extractable organic C (WEOC) were determined by DIMATOC 100 (DIMATEC, Germany) in the extracts.

The remaining soil was stored at −20°C until soil DNA was extracted according to Griffiths et al. (2000). Gene abundances of functional genes related to nitrification [archaeal amoA (AOA) and bacterial amoA (AOB)] and denitrification (nirK and nirS) were measured by SybrGreen-based quantitative real-time PCR, using the 7300 Real-Time PCR System (Applied Biosystems, Germany) as described by Hai et al. (2009). Briefly, the reaction volumes consisted of 25 µl including 12.5 μl Power SYBR® Green PCR Master Mix (Applied Biosystems), 0.5 μl 3% BSA (Sigma-Aldrich, Germany), 0.625 μl DMSO (Sigma, Germany), 2 μl DNA template, 0.5 μl of each primer (10 µM, Table 1), and 8.375 μl 0.1% DEPC water. All samples and standard curves (serial plasmid dilutions from 101 to 106 gene copies µl−1) were performed in triplicates, and at least four negative controls without DNA template were run per plate. Amplification efficiencies of 81–86% (amoA AOA), 79–82% (amoA AOB), 85–90% (nirK), and 86–91% (nirS) were achieved. Copy numbers were related to 1 g of dry soil.

Table 1 Primers used for real-time PCR quantification of functional genes [archaeal amoA (AOA) and bacterial amoA (AOB), nirK, and nirS]
Full size table

Statistical analyses were carried out using SPSS 11.5. Prior to analysis, data were checked for normal distribution and homogeneity of variances by the Kolmogorov–Smirnov test and the Levene test, respectively. The effect of the freeze–thawing, respectively, of the organic amendments at a given time point were tested on significance with the t test (P < 0.05).

Results and discussion

Due to the low N input in the straw treatment, soil NH +4 and NO −3 concentrations did not differ between the samples treated by freezing and thawing and the control samples constantly kept at 10°C. The measured concentrations remained low (3.3–5.5 µg NO −3 –N and up to 0.3 µg NH +4 –N g−1, respectively) and did not change during the investigation period (Fig. 1a, b). Similar observations were made for the soil WEOC contents in this treatment (28.3–54.2 µg C g−1; Fig. 1c). In contrast, addition of the lucerne-clover-grass-mix significantly increased the soil NO −3 –N (56.5–113.6 µg g−1), NH +4 –N (54.0–70.6 µg g−1), and WEOC (126.3–195.8 µg g−1) concentrations compared to the straw amendment (Fig. 1a–c). In this treatment, freezing and thawing significantly increased the amount of NH +4 –N and WEOC and led to lower NO −3 –N concentrations compared to the non-frozen control samples.

Fig. 1
figure 1

Amounts of ammonium, nitrate, and water-extractable organic carbon in the differently treated soils during incubation. Samples were taken at day 0 (day before freezing), day 4 (first day after thawing), day 6 (third day after thawing), and day 10 (1 week after thawing). Error bars indicate standard deviation of mean (n = 4). Different letters display at a given time point a significant difference between the freeze–thaw treatment and the unfrozen control treatment in the straw-amended (x/y) and the lucerne-clover-grass-mix-amended soil (X/Y) or a significant difference between the straw amendment and the lucerne-clover-grass-mix amendment in the unfrozen control treatment (a/b) and the freeze–thaw treatment (A/B)

Full size image

It is generally accepted that the amount of available inorganic N is one of the factors regulating denitrification rates in soil (Phillips 2008). Our results support our hypothesis that this is not only true for the vegetation period but can be extended to the winter period. Clark et al. (2009) could show that nitrification was the cause of NO −3 accumulation in frozen soils, and this process was consequently limited by NH +4 availability. Therefore, higher NH +4 concentrations are often associated with increased NO −3 concentrations (Elliott and Henry 2009).

High amounts of NH +4 during mid-winter thawing periods might be a result of increased soil protease activity, which is also related to the availability of C (Mrkonjic Fuka et al. 2009). In our study, we found higher concentrations of WEOC in the samples amended with the lucerne-clover-grass-mix (Fig. 1c) and correlating higher NH +4 concentrations. Moreover, WEOC determines denitrification activities (Phillips 2008). Thus, WEOC affects N turnover in two ways: on the one hand it stimulates the protease activity and the formation of inorganic N, and on the other hand it is a prerequisite for denitrification.

According to the different amounts of soil inorganic N and readily available organic C in the two treatments, higher abundances of genes involved in nitrification (AOA 0.8–1.6 × 106 and AOB 0.2–1.1 × 107 amoA copies g−1; Fig. 2a, b) and denitrification (nirK 2.0–2.5 × 108 and nirS 0.4–1.6 × 107 copies g−1; Fig. 2c, d) were determined in the soil amended with lucerne-clover-grass-mix. Interestingly, somewhat higher archaeal amoA (AOA; 4.7–9.7 × 105 copies g−1) than bacterial amoA (AOB; 1.8–5.8 × 105 copies g−1) gene copies were found in the straw-amended samples, whereas lucerne-clover-grass-mix-amended samples showed the opposite behavior (Fig. 2a, b). However, the influence of freezing and thawing resulted in both treatments in increased AOA/AOB ratios compared to the non-frozen control soil (Fig. 2b).

Fig. 2
figure 2

Abundances of the functional genes amoA (AOA and AOB), nirK, and nirS expressed as copy numbers per gram of dry soil in the differently treated soils during incubation. Samples were taken at day 0 (day before freezing), day 4 (first day after thawing), day 6 (third day after thawing), and day 10 (1 week after thawing). Error bars indicate standard deviation of mean (n = 4). For explanation of the different letters, please refer to Fig. 1. Ratios of AOA to AOB amoA copies are shown in boxes in Fig. 1b

Full size image

The result that AOB are more affected by the freeze–thaw event compared to AOA also supports the hypotheses by Schleper et al. (2005) and Valentine (2007) who presumed that archaea are more tolerant to stress conditions than bacteria. Several studies also showed that AOA/AOB ratios increase with decreasing availability of C and N (Leininger et al. 2006). It was also speculated that AOA might be more closely attached to soil particles which might result in a certain protection of AOA compared to AOB.

Concerning the genes involved in denitrification, the straw-amended samples also revealed lower copy numbers of both, nirK (0.4–1.5 × 108 copies g−1) and nirS (1.4–4.6 × 106 copies g−1) genes (Fig. 2c, d). However, both treatments exhibited considerably more nirK than nirS gene copies and showed decreased copy numbers after 1 day of thawing compared to the non-frozen control.

The higher nirK copy numbers could reflect a possible niche differentiation for microbes harboring one of the two functionally redundant nitrite reductases, where nirK-possessing bacteria could preferentially colonize habitats with more available C and N. This trend has also been observed when hotspots in soil exhibiting high amounts of available C and N have been compared with low activity areas (e.g., rhizosphere vs. bulk soil) (Sharma et al. 2005). In the legume mix-amended samples, a fast recovery of nirK copy numbers was detected, which was also found by Sharma et al. (2006), whereas the straw-amended soil showed a decrease in gene copies also in the control samples during the incubation period.

Our results indicate that the amount and quality of applied N have distinct effects on the abundance of genes involved in N cycling and the corresponding N transformations during freezing and thawing events. This is in line with previous studies showing the close link between the different transformation processes of the nitrogen cycle (Nannipieri and Paul 2009). This study indicates that the nitrogen stored in legumes, which are often used as green manure in agriculture, can fast be mineralized, nitrified, and denitrified during winter time as a result of freezing and thawing cycles. It could be considered if nitrification inhibitors applied after thawing might help to reduce transformations of the formed NH +4 into NO −3 and thus reduce N losses via leaching of NO −3 and emission of N2O and N2 by denitrification.