Fe(II)-OM Complexes Formed by Straw Returning Combined with Optimized Nitrogen Fertilizer Could Be Beneficial to Nitrogen Storage in Saline-Alkaline Paddy Soils

Abstract

Soil organic carbon (SOC) plays a crucial role in controlling the nitrate-dependent Fe(II) oxidation (NDFO) process, especially for saline-alkaline soils. The effects of straw returning combined with Nitrogen (N) fertilizer application on soil NO₃⁻-N content, Fe(II) form and nirK genes in saline-alkaline soil were studied in a five-year field experiment to explore the regulatory mechanism of SOC on NDFO process. Six treatments were designed with two factors (1) three straw returning rates (C0, C1 and C2, which was 0, 4500 and 9000 kg C ha⁻¹, respectively) and (2) two N fertilization rates (N1 and N2, which was 255 and 400 kg N ha⁻¹, respectively). Under both N levels, compared with C0 and C2 rates, NO₃⁻-N content was increased by 65% and 50% in C1 rate, respectively. NirK genes were decreased with straw returning, in which they were 42.9–58.8% lower in C1 and C2 treatments than that in C0 treatment, respectively. In the N1C1 treatment, the increase of SOC reduced the denitrification by converting aqueous Fe(II) (Fe(II)_aq) into Fe(II)-OM complexes and reducing the abundance of nirK genes. Overall, appropriate straw returning (C1) under optimal N fertilization rate (N1) could reduce N loss by decreasing the NDFO process in saline-alkaline paddy soils.

1. Introduction

N fertilizer has been used worldwide to improve soil fertility [1]. The amount of N fertilizer application in agro-ecosystems has been increased over the past few decades, especially in saline-alkaline land with low organic matter content and poor soil structure [2]. However, the excessive use of N fertilizer has caused environmental problems, such as N₂O emission [3]. Nitrate-dependent Fe(II) oxidation (NDFO) is considered as a main pathway of N loss in paddy soils, which occupies about 35% of N losses [4,5]. Under anaerobic conditions, NO₃⁻-N is successively reduced to NO₂⁻-N, NO, N₂O and finally to N₂ [6]. Therefore, it is of great significance to reduce the NO₃⁻-N loss by NDFO in saline-alkaline paddy soils.

The microbial-regulated NDFO requires the participation of obligate enzymes, such as NO₃⁻-N reductase (NR) and NO₂⁻-N reductase (NiR) [7]. NiR catalyzed the reduction of NO₂⁻-N to NO and N₂O, which was the key step of NDFO. NiR was encoded by nirK genes, which were suitable to research the denitrifying microorganisms in soil samples [7]. The structure and activity of denitrifying microorganisms determined the speed of denitrification and the generation of denitrifying products, which were affected by various environmental factors such as SOC, Fe, N, O₂ and so on [8,9].

The morphology of Fe(II) could also affect the occurrence of NDFO reaction [10]. Fe(II) exists in paddy soil as the free form of Fe(II)_aq, and also in the form of complexes such as Fe(II)-NOM and Fe(II)-OM. In recent years, studies have found that compared with Fe(II)_aq, Fe(II)-OM complex was more difficult to be oxidized by microorganisms [11]. The larger the ratio of SOC:Fe, the more stable the existing form of Fe(II) in soil, which was of great significance to the retention of N in paddy soil [12]. Previous studies have suggested that the organic amendments, such as straw and compost, contain higher levels of SOC, which could combine with Fe(II) to form Fe(II)-OM complexes, thereby increasing the Fe(II)-OM ratios in paddy soils [13].

In addition to increasing the proportion of Fe(II)-OM complexes, straw input also increased the proportion of macroaggregates and the stability of soil aggregates [14]. The distribution and combination of soil aggregates has important influence on soil properties, and different graded aggregates also show differences in physical and chemical properties and biological characteristics [15,16]. Jastrow et al. [17] found that SOC in macroaggregates was more tolerated to microbial decomposition, while SOC in small aggregates (such as microaggregates and silt+clay particles) was easily consumed by microorganisms. There were also differences in water distribution, air permeability and anaerobic characteristics of aggregates with different particle sizes [18]. Therefore, the composition of denitrifying microorganisms in aggregates with different particle sizes was different, and then led to the differences in the rate and product of NDFO in different aggregates. However, the relationship between NO₃⁻-N reduction and aggregate particle size was less studied.

Although the effect of carbon input on the NDFO process have been extensively studied, whether or not straw returning affects the process in saline-alkaline paddy soils remains unclear. In the Yellow River Delta, low SOC content and poor soil structure lead to serious N loss, in which NO₃⁻-N reduction is an important way that N is lost [6,19]. Straw returning combined with N fertilizer application could improve the transformation process of soil aggregates, and reduce N losses. Improving the utilization efficiency of N fertilizer ecology is of great significance to solve environmental problems. The main objectives of this study were to: (i) study the effects of Fe(II) form and content on NO₃⁻-N content under straw returning conditions; (ii) clarify the effects of straw returning on the proportion of macroaggregates and the changes of denitrifying enzyme (NR and NiR) activity in aggregates; (iii) investigate the regulation of soil C and Fe binding on nirK genes abundance under straw returning. We hypothesized that straw returning combined with N fertilizer decreased NO₃⁻-N reduction in NDFO by changing Fe(II) morphology, increasing the proportion of macroaggregates and decreasing the abundance of nirK genes.

2. Materials and Methods

2.1. Site and Soil Description

Beginning in 2016, a 5-year straw returning experiment was carried out in Kenli County, China (37°31 N, 118°32 E). The physicochemical properties of soil (0–20 cm) before the experiment were as follows: 0.29% salt content, pH 8.1, SOC 4.87 g kg⁻¹, total nitrogen (TN) 1.12 g kg⁻¹, available phosphorus (AP) 14.4 mg kg⁻¹, and available potassium (AK) 229 mg kg⁻¹.

2.2. Experimental Design

The experiment adopts two completely random factors of design: (1) two N levels (N1 and N2, which was 255 kg N ha⁻¹ and 400 kg N ha⁻¹, respectively) and (2) three straw returning levels (C0, C1 and C2, which was 0, 4500 and 9000 kg C ha⁻¹, respectively). The experimental plots were ploughed manually in early June each year to a depth of 20 cm. Rice straw (40% C) was cut into 5–8 cm long and applied into the plot before plowing. Urea (46% N) was used as N fertilizer. Base fertilizer (20%), early tillering stage (40%), tillering stage (20%) and flowering stage (20%) were applied four times in the growth period. The phosphorus fertilizer, superphosphate (16% P₂O₅, 128 kg P₂O₅ ha⁻¹), was applied as base fertilizers. Potassium fertilizer, KCl (50% K₂O, 229 kg K₂O ha⁻¹), was applied as the base fertilizer (50%) and the early stage (50%). Three repeated plots were set for each treatment, and the area of each plot was 20 m² (5 m × 4 m).

2.3. Soil Sampling

A soil auger with 2.5 diameter was used to collect soil samples (0–20 cm) in October 2021. Five soil samples were collected in each plot and mixed into one sample. Part of the samples was stored at −80 °C for extraction of microbial analysis. And the other samples were dried at room temperature to determine soil aggregates and soil physicochemical properties.

2.4. Soil Aggregates Fractions

The aggregates were separated by a wet-sieving method. Three aggregate fractions were obtained: macroaggregates (0.25–2.0 mm), microaggregates (0.053–0.25 mm) and silt+clay particles (<0.053 mm), respectively. Obtained soil aggregates were used to determine the activity of NR and NiR. Mean weight diameter (MWD) was used to indicate soil aggregate stability [20].

MWD was calculated as follows:

$$\mathrm{MWD}={{\displaystyle \sum }}_{i=1}^n{X}_i{W}_i$$

where $X_i$ is the average diameter of each particle size, $W_i$ is the proportion of soil aggregate mass in each particle size.

2.5. Soil Physicochemical Properties

Soil pH was measured by pH meter (Mettler Toledo FE 20, Shanghai, China). Dichromate oxidation digestion was used to determine soil SOC content [21]. TN was determined by the Kjeldahl method. The content of soil NO₃⁻-N was analyzed by a flow analyzer (AA3, Seal, Germany). Soil Fe(II)_aq and Fe(II)-OM content were determined by UV-Vis spectrophotometer (UV-6000PC, Shanghai, China) at 510 nm with 0.1% o-phenanthroline reagent. The activity of NR and Nir was determined by Hageman [22] methods.

2.6. Soil DNA Extractions and Sequencing

Depending on the manufacturer’s instructions for the PowerSoil^TM DNA Isolation Kit (MO BIO Laboratories, USA), the soil genomic DNA was extracted. Universal primer set of F338 (50-ACTCCTACGGGAGGCAGCA-30) and R806 (50-GGACTACVSGGGTATCTAAT-30) was used. PCR reactions used the following PCR procedure: initial denaturation at 95 °C for 3 min, 35 cycles at 94 °C denaturation for 30 s, primer annealing at 50 °C for 1 min, extension at 72 °C for 1 min; and a final extension of 10 min at 72 °C. The sequencing library was sequenced using Illumina MiSeq platform. Illumina MiSeq platform was used for microbial sequencing at the Majorbio Tech (Shanghai, China). Flash software was used to combine and reads to obtain the complete V3eV4 hypervariable region. Classification of OTUs were using the Ribosomal Database Project classifier. Chao1, Shannon and Simpson abundance-based indices were used to estimate the α-microbial biodiversity of the samples.

2.7. Quantitative PCR

Illumina MiSeq sequencing was used to determine community structure. Primer sets nirK876/nirK1070 (ATYGGCGGVCAYGGCGA/GCCTCGATCAGRTTRTGGTT) were used to amplify nirK genes using an Eppendorf Mastercycler. Extended range sequence cycling of nirK genes was performed at 98 °C for 2 min, followed by 30 cycles at 98 °C for 10 s, 58 °C for 30 s, and 72 °C for 45 s, and cycle at 72 °C for 10 min.

2.8. Statistical Analyses

Statistical analyses were conducted by using SPSS Statistics 20.0 (SPSS Inc., Chicago, IL, USA). Two-way ANOVA and Duncan test were used to analyze the variance of the data with different treatments, and p < 0.05 was considered to be statistically significant in our study. The relationship of TN, SOC, NO₃⁻-N, Fe(II)_aq and nirK genes was analyzed by Pearson correlation analysis. Structural equation models (SEM) were established using the Amos 23.0 software package. IBM SPSS Statistics 26 software was used for automatic linear modeling analysis. Origin Lab version 2018 was used for plots in this work.

3. Results

3.1. Soil Physiochemical Properties

Soil pH in N2 level was lower than in N1 level within the same straw addition amount (

Table 1. Results of two-way ANOVA for soil physiochemical properties.

Treatments		pH	SOC (g kg−1)	TN (g kg−1)	AK (mg kg−1)	AP (mg kg−1)
N1	C0	8.16 ± 0.01 Bb	7.91 ± 0.35 Aa	0.75 ± 0.05 Aa	74.48 ± 0.16 Aa	18.49 ± 0.63 Aa
	C1	8.07 ± 0.05 Aa	9.94 ± 0.80 Bb	1.09 ± 0.04 Bb	119.10 ± 1.47 Ac	29.64 ± 0.30 Ab
	C2	8.07 ± 0.03 Ba	10.16 ± 0.45 Bc	1.08 ± 0.08 Bb	90.99 ± 0.91 Ab	29.30 ± 0.25 Ab
N2	C0	8.12 ± 0.01 Bb	8.18 ± 0.60 Ba	0.76 ± 0.06 Aa	86.59 ± 1.45 Aa	22.16 ± 0.31 Ba
	C1	8.02 ± 0.02 Aa	8.67 ± 0.38 Ab	0.87 ± 0.10 Ab	129.93 ± 1.90 Ab	28.76 ± 0.22 Ab
	C2	8.04 ± 0.04 Aa	8.78 ± 0.42 Ac	0.88 ± 0.03 Ab	118.97 ± 0.89 Bb	27.57 ± 0.38 Ab
C		ns	<0.01	<0.01	<0.01	ns
N		<0.01	<0.01	<0.01	<0.01	<0.01
C × N		<0.01	<0.01	<0.01	ns	<0.05

Note: Data are shown as mean ± S.D. (standard deviation). The different lowercase letters indicate significant difference among straw input amounts of the same N application at p < 0.05. Different capital letters indicate a significant difference among N application amounts of the same straw input at p < 0.05.

), in which under the both N level, it was higher in C0 treatment than in C1 and C2 treatments. Straw returning increased the content of SOC under both N levels, in which it was increased by 16.53% and 15.73% in N1C1 and N1C2 treatments, respectively, compared with that in N2 level (Table 1). Compared with N2 level, TN content was increased by 25% in N1 level, and there was no significant difference between C1 and C2 treatments under both N levels.

Under both N levels, the content of soil NO₃⁻-N in C2 was significantly lower than in C1 (Figure 1a). Meanwhile, the content of Fe(II)_aq was the lowest in N1C1 treatment (3.34 mg kg⁻¹) and the highest in N1C0 treatment (5.14 mg kg⁻¹). The content of Fe(II)_aq in C1 and C2 treatments was lower than in C0 treatment after straw returning (Figure 1b). Besides, Fe(II)_aq content was negatively correlated with soil NO₃⁻-N content (Figure 1c). Straw returning significantly increased Fe(II)-OM content, in which it was 16.7~38.2% higher in C1 and C2 treatments than in C0 treatment. However, N fertilizer application was not significantly increased Fe(II)-OM content (Figure 1b).

3.2. Soil Aggregates

The proportion of soil aggregates was the highest in the macroaggregates fraction, followed by silt+clay particles and microaggregates, which accounted for 47.0–56.1%, 12.5–23.4% and 23.7–40.5%, respectively (

Table 2. Stability and size distribution of soil aggregates.

Treatments		Soil Aggregate Fractions (%)			MWD (mm)
		Macroaggregates	Microaggregates	Silt+Clay Particles
N1	C0	47.0 ± 1.8 Aa	12.5 ± 1.4 Bc	40.5 ± 3.6 Bc	0.5 ± 0.04 Aa
	C1	55.7 ± 2.1 Ac	19.2 ± 1.2 Aa	25.1 ± 1.1 Aa	0.7 ± 0.06 Bc
	C2	50.9 ± 1.8 Bb	23.4 ± 1.1 Ab	25.8 ± 3.4 Ab	0.6 ± 0.04 Bb
N2	C0	46.0 ± 2.3 Ba	18.0 ± 0.5 Ab	36.0 ± 2.8 Ac	0.5 ± 0.06 Ba
	C1	56.1 ± 1.9 Ab	20.2 ± 1.3 Aa	23.7 ± 0.8 Ba	0.7 ± 0.06 Ab
	C2	48.9 ± 1.5 Aa	23.0 ± 1.6 Bc	28.1 ± 2.8 Bb	0.6 ± 0.03 Aa
C		ns	<0.01	<0.05	<0.01
N		<0.01	<0.01	<0.01	ns
C × N		<0.01	<0.01	<0.01	<0.01

Note: Macroaggregates (0.25–2.0 mm), microaggregates (0.053–0.25 mm), silt and clay particles (<0.053 mm) and mean weight diameter (MWD). The different lowercase letters indicate significant difference among straw input amounts of the same N application at p < 0.05. Different capital letters indicate a significant difference among N application amounts of the same straw input at p < 0.05.

). Straw returning significantly increased the proportion of the macroaggregates fraction, which was increased by 8.7%, 3.9%, 10.1% and 2.9% in N1C1, N1C2, N2C1 and N2C2 treatments, respectively (p < 0.01). Compared with N1 level, the macroaggregates fraction of C0 and C2 treatments was higher than in N2 level.

The MWD was also significantly increased with straw returning: it was the highest in the C1 treatment, followed by C2 treatment, and C0 treatment was the lowest under both N levels (Table 2).

3.3. The Activity of NR and Nir in Soil Aggregates

Compared with C0 treatment, the activity of NR and Nir in C1 and C2 treatments was significantly decreased with straw returning under both N levels (Figure 2). Among the three different sizes of soil aggregates, the activity of NR and Nir was highest in silt+clay particles, followed by microaggregates and macroaggregates (Figure 2). Compared with N1 level, the activity of NR in N2C0 treatment was higher than that in N1C0 treatment, while no significant differences were found between C1 and C2 rates (Figure 2a). Within the same straw returning rate, the activity of Nir in macroaggregates fraction was significantly higher in N2 level than in N1 level, while there were no significant differences in C1 rate (Figure 2b). In silt+clay particles fraction, soil Nir activity was as high as 0.7mg g⁻¹ d⁻¹ in all treatments, except N1C1 treatment (Figure 2b).

3.4. The Abundance of NirK Genes

The abundance of nirK genes was increased with the increase of N application rate, in which it was the highest in C0N2 treatment, and the lowest was in C1N1 treatment (Figure 3). Compared with C0 rates, straw returning significantly decreased the copy numbers of nirK genes in C1 and C2 rates under both N levels (Figure 3). NirK genes copy numbers were negatively correlated with SOC, TN, NO₃⁻-N and MWD, while it was positively correlated with the content of soil Fe(II)_aq (Table S2).

3.5. Composition of Soil Microbial Community

At class level, the dominant microbial community in six treatments were Gammaproteobacteria, Anaerolineae, Vicinamibacteria, Alphaproteobactia and Actinobacteria (Figure 4 and Table S1). Under N1 level, compared with C1 treatment, the relative abundance of Gammaproteobacteria was decreased in C0 and C2 treatments, but the relative abundance of Anaerolineae and Actinobacteria was increased. Under N2 level, compared with C2 treatment, the relative abundance of Vicinamibacteria was increased in C0 and C1 treatments, while the Actinobacteria was decreased (Figure 4 and Table S1). With the increase of N application rate, the relative abundance of Actinobacteria showed a decreasing trend, whereas the relative abundance of Gemmatimonadetes, Gemm-5, Bacteroidia and Cytophagia was increased.

3.6. The Relationship between Soil NO₃⁻-N Content and Soil Properties

The direct and indirect influences of SOC, TN, MWD, Fe(II)_aq and the abundance of nirK genes on soil NO₃⁻-N content were inferred from SEM (Figure 5). SOC and TN were positively correlated with soil NO₃⁻-N content. While soil TN had a negative relationship with Fe(II)_aq, but had a positive relationship with the MWD. The abundance of nirK genes was positively driven by Fe(II)_aq, but negatively regulated by MWD.

4. Discussion

4.1. Effects of Straw and N Additions on Soil Properties

SOC was one of the main factors of soil quality. Straw returning could increase SOC content and enhance soil carbon sequestration ability in coastal paddy soils [19,23]. In the present research, 5 years of straw application significantly increased SOC content, which confirmed the role of straw in SOC storage (Table 1). Wei et al. [24] showed that SOC content was increased significantly with the increase of straw returning amount. Therefore, carbon input in farmland was the main source of SOC content [13]. Previous studies have found that N fertilizer could promote SOC accumulation by inhibiting soil respiration and decomposing carbon stability [25]. In this study, 5 years of N fertilizer application increased the content of SOC. While SOC content in N2 level was significantly lower than in N1 level (Table 1), which indicated that excessive N application could not increase SOC content.

In this study, soil TN content in C1 and C2 treatments was significantly higher than in C0 treatment (Table 1). Wei et al. [24] also confirmed that straw returning could increase the content of TN through a 4-year field positioning experiment. He et al. [26] showed that straw returning significantly increased soil TN and microbial N contents in 0–10 cm and 10–20 cm soil through a 17-year field positioning experiment. While soil TN content in N2 level was significantly lower than in N1 level (Table 1). This phenomenon might be due to denitrification loss of N fertilizer in anaerobic environment. Straw returning combined with N fertilizer could significantly affect the relative abundance of denitrifying microorganisms. Excessive N application provides sufficient N source for denitrifying microorganisms. Straw returning combined with N fertilizer significantly increased soil NO₃⁻-N content (Figure 1a), which was similar to the results of other research with organic amendments [27]. Furthermore, we found that NO₃⁻-N content in C1 treatment was higher than in C2 treatment under the same N level, which was possibly because the higher abundance of nirK genes in C2 treatment could lead to more NO₃⁻-N reduction (Figure 1a and Figure 3).

N in macroaggregates generally could not be mineralized, while in microaggregates, they were often subject to runoff and wind erosion, thus nutrients in microaggregates will be at greater risk of release. Straw returning significantly reduced the proportion of Silt+Clay fraction (<0.053 mm), and increased the proportion of macroaggregates (0.25–2 mm) (Table 2), which was because organic amendments combined soil particles to form macroaggregates. Straw returning also significantly increased soil MWD and improved the stability of soil aggregates (Table 2). Previous studies have found that straw decomposition could produce polysaccharide gum, fat and other intermediates, which were conducive to the formation of soil humus and other important organic cementing substances [28]. On the other hand, straw returning could improve microbial activity and promote the growth of fungal flora, and fungal mycelia from microbial secretions could also cement soil particle composite [17].

4.2. Responses of NO₃⁻-N to Different Forms of Fe(II), Soil Aggregates and NirK Genes

In this study, straw application significantly decreased the content of Fe(II) aq, but increased the content of Fe(II)-OM (Figure 1b). Previous studies have found that the oxidation rate of Fe(II) was decreased after OM and Fe(II) complexes to form Fe(II)-OM [29,30]. Peng et al. [11] found that Fe(II)-OM could inhibit the oxidation of Fe(II) by nitrate reducing bacteria, such as Acidovorax sp. strain BoFeN1, possibly because the complex changed the size of Fe(II)-OM and prevented it from entering the extracellular membrane. Most Fe(II) existed in paddy soil as OM-complexes, which directly inhibited denitrifying microorganisms in the process of NDFO. We believe that the change of Fe(II) was one of the important reasons for N retention in saline-alkaline paddy soils.

Denitrification mainly occurred in the anaerobic environment [4]. Compared with macroaggregates, small aggregates (such as microaggregates and silt+clay particles) have lower intergranular porosity and permeability, which were conducive to the formation of anaerobic environment and accelerated the denitrification in paddy soil [31]. Similar to previous studies, the results showed that the activity of NR and Nir was increased with the decrease of aggregate size (Figure 2). Denitrifying microorganisms were mostly heterotrophic microorganisms, which need SOC as nutrient substrate for their growth. Although SOC in small aggregates was more stable than that in macroaggregates, due to the certain spatial isolation between microorganisms and SOC in macroaggregates, SOC in macroaggregates was more difficult to be utilized by denitrifying microorganisms which was one of the reasons why denitrifying enzyme in small aggregate was higher than in macroaggregates [32].

Previous research has shown that these factors could affect denitrification, including TN, nirK genes, O₂, NO₃⁻-N content, SOC content and microbial community composition [33,34]. In this study, SEM showed that Fe(II)_aq and MWD mainly affected the content of NO₃⁻-N by directly affecting the abundance of nirK genes (Figure 5). And the differences in NO₃⁻-N were primarily explained by the corresponding differences in TN, SOC, Fe(II)_aq, MWD and nirK genes (Figure 5). Moreover, TN and SOC could also change the abundance of nirK genes (Table S2), which was negatively correlated with C/N in sediment. The reason might be most denitrifying bacteria were heterotrophic, so the content of C and N was very important in determining community structure [35]. In addition, soil C/N ratio in C0N1 and C0N2 treatments was relatively low, resulting in more mineral N released and faster decomposition of SOC, which was conducive to the growth of denitrifying bacteria [19]. In addition to the abundance of nirK genes, the composition of the microbial community was also an important factor affecting denitrification. In this study, there was no significant difference in microbial community composition among different treatments, which indicated that there was no correlation between the composition of denitrifying microorganisms and NO₃⁻-N content (Figure 1a and Figure 4). The main reason might be the repeated disturbance of topsoil by continuous cultivation and the continuous renewal of soil aggregates [36].

5. Conclusions

Due to the 5-year straw returning combined with N fertilizer application, the content of SOC and TN in paddy soil was increased, in which it was highest in C1N1 treatment. The increase of SOC and TN content promoted the transformation of Fe(II)_aq to Fe(II)-OM complexes in paddy soil. Meanwhile, straw returning was beneficial to the formation and stability of macroaggregates. Compared with C0 treatments, straw returning (C1 and C2 treatments) decreased the activity of NR and NiR in aggregates, and then decreased the abundance of nirK genes. The decrease of Fe(II)aq content and nirK gene abundance inhibited the NDFO process and reduced N loss in paddy soil. In this study, straw returning could increase the content of NO₃⁻-N, and the NO₃⁻-N retention capacity was higher in N1C1 treatment. Therefore, strategies of lower straw returning (C1) under optimal N fertilization rate (N1) could improve soil aggregate structure and NO₃⁻-N retention capacity, which reduced the N loss in coastal saline-alkaline soils.