The Reduction of Nitrogen Fertilizer Rate Shifted Soil Bacterial Community Structure in Rice Paddies

Abstract

In order to achieve reasonable yield while keeping environmental risks low, nitrogen fertilizer reduction has been adopted for in rice cultivation. The response of the soil microbial community structure to this management is not fully understood. In this study, the treatments comprising conventional farming practices (330 kg ha⁻¹), reduced N application (270 kg ha⁻¹ and 300 kg ha⁻¹, respectively), and a control without N application were set up in order to reveal the effects of N application rate on the soil microbial community composition in rice paddies. It was discovered that Proteobacteria, Acidobacteria, Actinobacteria, and Chloroflexi represented the most abundant bacterial phyla in all samples. The assembly of the soil bacterial community differed among the treatments, with NH₄⁺-N, available phosphorus (AP), and organic matter (OM) as key drivers. The reduction of N application by 20% decreased soil NO₃⁻ up to 32% and increased the abundance of the total functional pathways, especially those associated with carbon fixation, N, S, and CH₄ metabolism, whereas N reduction by 10% increase soil N accumulation and soil bacterial richness. In summary, a reduction of N fertilizer by up to 20% compared to the amount used in traditional practices could most effectively regulate the soil bacterial community composition and increase the predicted functional groups associated with N transformation, while maintaining lower soil nitrogen contents.

1. Introduction

Rice is a staple food for half of world’s population, and it plays a crucial role in Chinese agriculture, with its rice yield accounting for almost half of the world’s total production. Remarkable progress has been made in research regarding the productivity of the rice field ecosystem, and super high yields have been achieved using different agricultural management techniques [1]. One of these strategies is to increase N input, which is known to be critical for the increased yield of cereal crops [2]. However, the utilization rate of nitrogen fertilizer in Chinese rice production is only 30%~35%, which is lower than the world average of 46%. The average amount of nitrogen applied in rice field production in the Yangtze Delta in China reached 250 kg hm⁻², far exceeding the world average, at more than 1.2–1.6 times the recommended rate [3]. It is even recommended to use even less N for the aim of profit and ecosystem sustainability [2].

Nitrogen application holds significant ecological importance, and there is increasing attention paid to its impact on the global nitrogen cycle. The inefficiency of nitrogen fertilizer use, along with issues such as the deterioration of soils, nitrogen loss through leaching, N₂O emissions, and the imbalance of soil microbial communities, have become focal points [1,4,5,6,7]. Additionally, the fluctuating dry–wet cycles affect the redox potential of paddy soil, leading to violent fluctuations in elements that impact microbial-driven bio-transforming processes [8,9,10].

The soil bacterial community is an important part of the soil ecosystem, which can decompose and transform soil organic matter, regulate the soil carbon and nitrogen cycles, and promote the soil biogeochemical process [11]. Nitrogen is the basic element regulating the composition and function of the soil bacterial community [12]. The reasonable application of nitrogen fertilizer can increase the nutrient availability of soil and change the physical and chemical properties of the soil, as well as the structure and function of the soil microbial community [13]. These changes, in turn, can promote plant growth and ensure plant health by accelerating soil nutrient circulation, inhibiting soil pathogens, and reducing disease incidence [11,12,13]. The diversity and richness of the soil microbial community structure is the basis of the soil nutrient cycle [14].

Soil microorganisms regulate N transformation in many ways, and the key microbial groups and essential functional genes involved in N cycling have been studied extensively [6,15]. A diversity of soil microorganisms, with the function of N-fixation, nitrification, denitrification, and ammonia oxidation, among others, and their succession may have a significant impact on the transformation of soil nitrogen and eventually, on the nitrogen use efficiency in regards to rice growth [16,17]. Studies have shown that soil microbes may be impacted by fertilization and other farming techniques [18], with contradictory results [19,20]. It is also suggested that biological attributes are more representative of agricultural land conditions than are physicochemical variables [21,22]. However, there is still a lack of understanding regarding how fertilizer management affects the responses of the soil microbial community and functional groups involved in N activities.

Previous studies have demonstrated that in a reasonable range of N application, an increase in grain yield can be achieved, accompanied by an acceptable nitrogen use efficiency (NUE) [23]. This needs to be further elucidated in regards to the effects of N application on soil bacterial community composition and soil properties in rice cultivation This paper aims to reveal the following: (1) the effects of N fertilizer reduction on the major soil properties in a super rice production system; (2) the shift in soil bacterial community structure and the functional groups as a result of N reduction, along with the direct influencing factors. The results are expected to provide a scientific reference for N use regulation in sustainable agriculture.

2. Materials and Methods

2.1. Rice Plot Experimental Design

A field experiment was carried out at an experimental site in Yangzhou City, Jiangsu Province, China. This site belongs to the warm temperate monsoon climate zone, with an average annual temperature of 14.1 °C. The average annual precipitation is 892.3 mm. Wheat–rice rotation is the typical agricultural mode in this region. Wheat was the pre-crop, and the soil type was sandy loam. Organic matter (OM) (28.7 g kg⁻¹), alkali-hydrolyzable nitrogen (108 mg kg⁻¹), available phosphorus (A–P) (34.5 mg kg⁻¹), and available potassium (126 mg kg⁻¹) were all present in the soil.

The experiment was carried out using the completely randomized block arrangement (Figure S1), with the plot size of 5 × 8 m. The following treatments were adopted in the experiment: a blank control with no N fertilizer application (CK), a traditional farming management (TN) system, with a N application rate of 330 kg ha⁻¹ (in the form of urea), two treatments with reduced nitrogen application, i.e., low nitrogen (LN), with a N dosage of 270 kg ha⁻¹, and high nitrogen (HN), with a N dosage of 300 kg ha⁻¹. For the treatments with N fertilizers, 60% of the N was applied as basal fertilizer, and 40% was used for topdressing. For all treatments, CaP₂H₄O₈ and KCl were applied as basal fertilizers at the rates of 745 kg ha⁻¹ and 240 kg ha⁻¹, respectively.

The rice cultivar super rice (Oryza sativa L.) Ningjing 4 was used. The rice seedlings were transplanted on 20 June 2023. Three rice seedlings were planted per hill, at a density of 10 cm in width and 25 cm in length. Three replicate plots were set up for each treatment. The other agricultural managements were performed following traditional farming practices.

2.2. Soil Sampling

Soil samples were taken for analysis at the harvest stages of rice (10 November 2023). A total of 15 soil cores (diameter of 3 cm) were randomly taken from each plot, following a “S” route method. The surface soil from 0 and 20 cm in depth was collected with a auger, cleaned with ethanol, thoroughly mixed in a sterile zippered bag, and divided into two portions. One was transferred to a sterile falcon tube, pre-frozen at −80 °C, then freeze-dried, according to the manufacturer’s instructions (FD-1A-50, Jiangsu Tianling Instrument Co., Ltd., Yancheng, China), for molecular analysis, while the other part was air dried for the examination of the physiochemical properties of the soil.

2.3. Determination of Soil Physiochemical Parameters

To determine the physiochemical properties of the soil, the air-dried soils were thoroughly ground and sieved using a 0.15 mm mesh sieve. To determine the pH and electric conductivity (EC), a soil sample was diluted in DI water (1:2.5, m:v) using Leici pH and EC meters (Shanghai INESA Scientific Instrument Co., Ltd., Shanghai, China), respectively. The basic parameters, including soil organic matter, soil nitrate (NO₃⁻-N), ammonium (NH₄⁺-N), alkali-hydrolyzable N, available phosphorus, and potassium were also determined following the protocol described by Bao et al. [24]. In short, the K₂Cr₂O₇ external heating method was used to quantify soil organic matter; spectrophotometry was used to measure soil nitrate-N and ammonium-N. The alkaline hydrolysis diffusion method was used to quantify alkali-hydrolyzable nitrogen. Soil available potassium was measured using flame photometry, and soil available phosphorus was measured using the Mo-Sb colorimetric method.

ANOVA was employed, using the statistical program SPSS (26.0), to determine whether there were differences in the soil physiochemical parameters among the treatments.

2.4. Soil Microbial Community Analyses

Using the PowerSoil DNA Isolation Kit (QIAGEN Inc., Valencia, CA, USA), the complete soil bacterial DNA was recovered. By using bacterial primer sets 338F and 806R [25], the hypervariable V3–V4 regions of 16S rDNA were then amplified following the amplification conditions, as previously described [26]. The qualified PCR products were used for Illumina sequencing at Majorbio Bio-Pharm Technology Co. Ltd. (Shanghai, China) using the Illumina MiSeq platform (Miseq PE300, Illumina, San Diego, CA, USA).

The sequence information on the bacterial community was analyzed using the Majorbio I-Sanger Cloud Platform (http://www.i-sanger.com, accessed on 27 May 2024). The qualified sequences after filtration were aligned against the SILVA132 database (https://www.arb-silva.de/, accessed on 30 May 2024), and the taxonomy details of the sequences were obtained. The operational taxonomic units (OTUs) were assigned based on 97% similarity [27]. Bacterial diversity estimator indices, including the ACE, Shannon, and Sobs indices, were calculated based on the OTU distribution after the number of reads in each sample was normalized.

The alpha diversity indices of the bacterial taxa were compared between treatments using one-way ANOVA with the Tukey–Kramer post hoc test. Non-metric multidimensional scaling (NMDS) and ANOSIM analyses were performed to estimate the beta diversity [28]. Following that, the study employed redundancy analysis (RDA) to investigate the impact of soil properties on the composition variation among bacterial communities. Based on the Spearman correlation, the relationship between the dominating bacterial taxa and soil properties was computed and displayed as a heatmap.

The FAPROTAX [29] analysis method was utilized to predict the potential functions of the bacterial communities present in the soil samples, based on which, the sequences were classified into different functional groups, including C, N, S, P transformations, and other metabolic mechanisms. The quantity and proportion of various functional groups were obtained, especially those engaged in various N transformation. One-way ANOVA was used to compare the variations in the proportion of functional groups among the treatments.

3. Results

3.1. Effects of N Application on Soil Physiochemical Parameters

The major soil properties from different treatments are listed in

Table 1. Major soil chemical properties in treatments with different nitrogen fertilizer rates.

Treatments	pH	EC μS cm−1	AN mg kg−1	NN mg kg−1	KN mg kg−1	AP mg kg−1	AK mg kg−1	OM g kg−1
CK	6.67 ± 0.07	92.8 ± 4.9 b	5.02 ± 0.41 c	3.46 ± 0.15 c	115 ± 3.9 c	88.8 ± 7.0 a	149 ± 4.0 a	21.3 ± 1.9 a
TN	6.73 ± 0.13	104.5 ± 7.1 ab	6.55 ± 0.41 ab	5.54 ± 0.12 b	140.7 ± 2.2 b	78.3 ± 5.6 b	126.7 ± 7.3 b	18.1 ± 1.6 b
LN	6.65 ± 0.07	112.9 ± 8.8 a	5.85 ± 0.62 bc	3.77 ± 0.63 c	133.7 ± 2.9 b	77.3 ± 1.1 b	137.1 ± 11.7 ab	20.2 ± 1.5 ab
HN	6.783 ± 0.06	88 ± 8.7 b	7.44 ± 0.55 a	8.19 ± 0.18 a	160 ± 12 a	77.9 ± 3.1 b	133.6 ± 9.6 ab	18.5 ± 0.9 b

Lowercase letters indicate significant differences among the treatment within one column (LSD, p < 0.05). EC: electrical conductivity; AN: NH₄⁺-N; NN: NO₃⁻-N; KN: alkali-hydrolyzable nitrogen; AP: available p; AK: available K; OM: soil organic matter.

. Not surprisingly, the lowest content of all forms of nitrogen (including ammonium, nitrate, and alkali-hydrolyzable nitrogen) was detected in the treatment without N application (CK), followed by that with low nitrogen input (LN), with the highest observed in the treatment employing a high nitrogen rate (HN) (p < 0.05). In comparison to traditional farming N applications, the practice of nitrogen fertilizer reduction decreased the amount of different forms of nitrogen in the soil. Soil NH₄⁺-N, NO₃⁻-N, and alkali-hydrolyzable nitrogen decreased by 10.7%, 31.9%, and 5%, respectively. The N fertilizer reduction practice increased soil EC significantly when compared with that of other treatments. The soil in the control treatment displayed significantly higher concentrations of AP, AK, and OM (p < 0.05).

3.2. Effects of N Application on Bacterial Community Structure

Using Illumina sequencing, a total of 1,372,015 qualified sequences, with an average length of 428 base pairs, were retrieved from 24 soil samples. The diversity indices, including the Chao, ACE, Shannon, and coverage, represent the richness and evenness of the soil bacterial community.

Table 2. The soil bacterial alpha indices for different treatments.

Treatments	Chao	ACE	Shannon	Coverage
CK	5256.5 ±754.49	5516.6 ± 1328.4	6.71 ± 0.19	0.963
TN	5131 ± 519.9	5123.3 ± 530.35	6.83 ± 0.16	0.965
LN	5214.8 ± 371.76	5130.5 ± 349.3	6.82 ± 0.05	0.965
HN	5560.9 ± 233.72	5574.3 ± 250.7	6.99 ± 0.1	0.961

lists the alpha diversity indices of the soil bacterial community calculated from the treatments. Overall, the treatment with the 300 kg ha⁻¹ N application (HN) exhibited the highest Chao, ACE, and Shannon indices, indicating a high level of soil bacterial diversity and abundance in this treatment. Nevertheless, there was no statistically significant variation between the treatments.

The study employed non-metric multidimensional scaling analysis (NMDS) to create a figure that illustrates the differences in microbial communities across the various treatments. Nitrogen fertilizer management shifted the composition of the soil bacterial population (Figure 1), and the control sample without N application was clearly distinguished from the others. There was a significant variation among replicates of the traditional agricultural management system (TN), which was still different from the treatments with lower (LN) or higher N (HN) use. This indicates that the N application rate had a great impact on the soil bacterial community composition. The results of the ANOSIM test additionally indicated that there was a significant difference (p < 0.05) in the compositions of the soil bacterial communities among the treatments.

3.3. Soil Bacterial Community Composition

In total, 48 phyla and 1132 genera were obtained by the taxonomic analysis of the high-quality data from Illumina sequencing across all treatments. Figure 2 displays the taxonomic composition of the bacterial sequences at the phylum and genus levels. Proteobacteria was the most common phylum in all treatments, making up an average of 27.8% of all reads (Figure 2A). Chloroflexi (20.8%), Acidobacteria (16.4%), Actinobacteria (16.1%), Firmucates (5.01%), and Gemmatimonadetes (3.29%) were the next most abundant phyla. These first six phyla comprised 89.38% of the total bacterial abundance. With a proportion of 31.74% of all sequences, the largest percentage of Proteobacteria was found in treatment TN, or the traditional farming treatment, whereas the lowest ratio, 24.48%, was found in treatment HN, which used more nitrogen fertilizer. The proportions of each bacterial phylum in the samples were not statistically significantly different.

In Figure 2B, the sequences percentage > 1% of the bacterial taxa that were relatively numerous were listed. g__norank_f__norank_o__norank_c__Subgroup_6, g__norank_f__Anaerolineaceae, and Arthrobacter, which are members of the phylum Acidobacteria, Chloroflexi, and Actinobacteria, respectively, were the most common bacterial genera. With ratios of 6.52%, 4.56%, and 4.5%, respectively, these three genera displayed the highest relative abundance under the LN (reduced nitrogen application) treatment. The relative abundance of the dominating taxa (>1%) varies little between treatments. However, a significant difference was observed only in the bacterial phyla with relative abundance less than 1%, which are often classified as rare taxa.

To identify the bacterial biomarker taxa in the paddy soils of different treatments, a linear discriminant analysis effect size (LEfSe) assessment was performed. The results demonstrated that significant bacterial taxa changed at the class level (Figure 3A) in the treatments. There were considerably higher bacterial biomarkers in CK than in other treatments (Figure 3B), with 12 and 14 taxa with LDA scores of 2~3 and 3~4, respectively. The treatment with the second most abundant biomarkers was the traditional agriculture (TN) treatment, followed by the increase N treatment. Reduced nitrogen treatment displayed the lowest number of biomarkers.

3.4. The Effects of Soil Properties on Bacterial Communities

Redundancy analysis (RDA) was used to examine the relationship between the chemical characteristics of the soils and the distribution of the soil bacterial community structure (at the phylum level) in samples from different treatments (Figure 4). The first and second axes explained 48.05% of the total variance (RDA1 = 35.61% and RDA2 = 12.44%). The results demonstrated that the bacterial community structure in the traditional N application treatment was more affected by EC and soil ammonium-N content, whereas the increased nitrogen treatment was more influenced by soil pH and nitrate-N. The results are more complicated for the CK and LN treatments, as there was a significant variation among the replicates. Generally, soil organic matter concentration (r² = 0.61, p = 0.01), available phosphorus (r² = 0.51, p = 0.04), and NH₄⁺-N content (r² = 0.48, p = 0.04) were shown to be the main environmental factors influencing changes in the structure of the soil bacterial population.

To further study the effects of the soil physiochemical parameters on the relative abundance of prominent bacterial taxa (the top 10 phyla), the Spearman rank correlation approach was adopted, with the results listed in

Table 3. The correlation between the relative abundance of the major bacterial phyla and soil physiochemical parameters, as calculated by Spearman correlation analysis.

Phylum	NN	AN	KN	AP	AK	EC	pH	OM
Proteobacteria	0.09	0.39	−0.06	0.01	−0.02	0.52	−0.01	−0.37
Chloroflexi	0.08	−0.05	0.24	−0.17	0.20	−0.32	0.06	−0.46
Acidobacteria	0.24	0.58 *	0.50	−0.52	−0.41	0.06	0.29	−0.51
Actinobacteria	−0.45	−0.08	−0.43	0.19	0.23	0.24	−0.37	0.22
Firmicutes	−0.36	−0.13	−0.33	0.25	0.33	0.12	−0.40	0.33
Gemmatimonadetes	0.19	0.55	0.44	−0.80 **	−0.69 *	0.46	0.03	−0.33
Bacteroidetes	−0.10	−0.56	−0.17	0.23	0.33	−0.53	0.07	0.17
Rokubacteria	0.52	0.81 **	0.59 *	−0.48	−0.39	0.22	0.42	−0.58 *
Nitrospirae	0.43	0.01	0.30	0.00	0.13	−0.06	0.09	−0.43
Latescibacteria	0.57	0.50	0.83 ***	−0.76 **	−0.65 *	−0.15	0.50	−0.57

NN: nitrate-N; AN: ammonium-N; KN: alkali-hydrolyzable nitrogen; AP: available p; AK: available K; EC: electrical conductivity; OM: soil organic matter. * represent a significant difference between the treatments (* 0.01 < p ≤ 0.05, ** 0.001 < p ≤ 0.01, *** p ≤ 0.001).

. There was a significant positive correlation between the relative abundance of Acidobacteria and Rokubacteria and the soil ammonium nitrogen content (AN), while there was a negative correlation between the relative abundances of Gemmatimonadetes and Latescibacteria and the soil’s available phosphorus (AP) and potasium (AK) contents. The phylum Rokubacteria was adversely affected by the organic matter content (OM) of the soil.

3.5. Functional Prediction of Microbial Groups

PICRUSt analysis was performed to predict the potential functions of the soil bacterial communities in the different treatments (Figure 5). The dominant functional pathway abundance varied among the treatments. Interestingly, the pathways involved in cellular processes, environmental information processing, genetic information processing, and metabolism were the most enriched in the LN treatment. As compared to the control, the reduced nitrogen treatment increased the soil functional groups by 6.22%, on average. Also, the HN treatment showed higher abundances for all functional pathways, except for cell motility and signaling molecules and interaction, than did TN (Figure 5A). Further analyses on the energy metabolism pathways revealed that the predicted pathways associated with C fixation, methane production, and oxidative phosphorylation were relatively more abundant (Figure 5B). Similarly, the two treatments with reduced nitrogen application exhibited abundances of associated pathways for C fixation, CH₄ metabolism, and oxidative phosphorylation. LN also showed a higher abundance of N and S metabolism pathways. Overall, N fertilizer reduction was beneficial for cell growth and metabolism, with the best results observed when ~20% N was reduced.

4. Discussions

4.1. Soil Physicochemical Parameters as Influenced by N Fertilizer Reduction

The two main inorganic nitrogen sources in the soil, ammonium (NH₄⁺) and nitrate (NO₃⁻), are essential for plant uptake. The efficiency of using nitrogen is significantly impacted by the conversion of nitrogen [30]. Previous results have shown that the reduction of N input increased N recovery efficiency and greatly reduced environmental risks [31]. This is in accordance with our result when 20% N fertilizer reduction was applied. Nevertheless, this experiment showed that when the N use rate was reduced by approximately 10%, rice soil nitrate, ammonium, and alkali-hydrolyzable N all increased. In this study, too much nitrogen input (TN) might have caused vigorous growth of the rice plants, accompanied by rapid nutrient uptake [32], leaving lower N accumulation in soils. Furthermore, using excess N may extend the nutritional growth and postpone rice maturity, which is detrimental to grain development [33]. Indeed, the rice yield analysis showed that, with a N fertilizer reduction, LN achieved a comparable yield to that of TN (Table S1), suggesting a high nitrogen use efficiency of the N fertilizer reduction treatments. Other soil nutrients, including soil available P and K, also declined in the traditional farming treatment, most likely as a result of the rapid growth of the rice plants [34].

4.2. Shifts in Bacterial Diversity and Community Composition

Fertilization techniques change the community makeup of the soil microorganisms by providing them with nutrients and energy for growth, which in turn mediates the soil biogeochemical cycles [35,36,37]. Similarly, this study demonstrated that while reducing 20% nitrogen decreased soil bacterial richness, reducing 10%r N improved both soil bacterial richness and evenness. The stimulation of copiotrophic bacteria, which use the K strategy for growth [38], was the reason for the enhanced bacterial richness that we also noticed in the previous study [26].

The NMDS study showed that the composition of the soil bacterial population varied significantly between the treatments. It is not surprising that samples without N treatment (CK) were distinct from the others, as N input may promote the proliferation of soil microorganisms [15]. Previous research showed a strong correlation between the N fertilizer application and the composition of the microbial population [39].

Proteobacteria, Chloroflexi, Acidobacteria, and Actinobacteria were found to be the most common bacterial taxa in all treatments. Proteobacteria and Chloroflexi are known to be highly prevalent in rice fields [40] and to be largely linked to nitrogen turnover [15]. According to Bose et al. [41], Acidobacteria may also play a significant role in the conversion of NO₃⁻ to NH₄ ⁺ in rice paddies. This is supported by the positive association found in Table 3 between bacterial relative abundance and NH₄⁺ content. These were consistent with the results of the earlier study, which found that N fertilization and the proportions of Actinobacteria and Proteobacteria—also referred to as copitrophic bacteria—increased [42].

The bacterial community structure of the soil was significantly impacted by the NH₄⁺-N, A-P, and OM contents of the soil. The prospective nitrate-reducing microorganisms Acidobacteria and Rokubacteria showed a positive correlation with soil NH₄⁺-N [41], but Gemmatimonadetes and Latescibacteria showed a positive correlation (p < 0.05) with AP, indicating that these microbes may be engaged in P cycling.

Microorganisms participating in many processes, including N fixation, ammonium and nitrite oxidation, denitrification, etc., were responsible for driving N transformation in soils [15]. Considering the bacterial functions, the LN treatment had the highest abundances of pathways detected, specifically those associated with carbon fixation, N, S, and CH₄ metabolism. Thus, in order to improve our understanding of the function of soil ecosystems after the reduction of nitrogen fertilizer, future research should concentrate on both the impacts of a reduction of N fertilizer application on the phylogenetic diversity of bacteria, as well as the functional diversity of bacteria. Further, FAPROTAX analysis [43] was used to forecast the functional groups. It was shown that, except for the nitrification process, which was noticeably elevated in soil samples of the traditional farming N treatment, there was very little variation in the bacterial groups involved in N transformation, suggesting that the treatment was undergoing vigorous nitrification.

5. Conclusions

This study examined the soil characteristics and the dynamics in the soil bacterial population at rice harvest under reduced nitrogen treatment. It was discovered that when compared to the farmers’ traditional nitrogen application rate, a 10% N fertilizer reduction improved the contents of various forms of N, as well as the bacterial richness of the soil. Proteobacteria, Acidobacteria, Actinobacteria, and Chloroflexi were the most common bacterial phyla observed throughout the treatments; only the phylum Acidobacteria showed a positive connection with soil NH₄⁺ content. Soil NH₄⁺-N, available P, and OM were the key soil parameters that influenced the assembly of the bacterial community composition. LN treatment had the highest pathway abundance, along with significantly higher ratios of the N transformation groups than noted for the control. In summary, reducing the application of N fertilizer up to 20% in comparison to amounts used in traditional practices could most effectively regulate the soil bacterial community composition and increase the predicted functional groups associated with N transformation, while maintaining lower soil nitrogen contents.