Exploring Suitable Nitrification Inhibitor in an Intensively Cultivated Greenhouse Soil and Its Effect on the Abundance and Community of Soil Ammonia Oxidizers

Abstract

The application of nitrification inhibitors (NIs) is an effective way to reduce soil nitrogen (N) losses and increase crop N uptake. Yet, the efficacy of NIs commonly varies with dosages, crop systems and soil environmental conditions. Hence, clarifying the suitable type and dosage of NIs is extremely important for structuring the best N management regime at a regional scale. Here, based on microcosm experiments, we evaluated the influence of three widely used NIs [Dicyandiamide, DCD; 3,4-Dimethylpyrazole phosphate, DMPP; 2-chloro-6-(trichloromethyl) pyridine, Nitrapyrin] on the nitrification activity of an intensively cultivated greenhouse soil. The results showed that both DCD and DMPP imposed a transient inhibition on nitrification (less than five days) regardless of the dosages applied, and, on the contrary, Nitrapyrin presented a persistent suppression, with a longer duration of the inhibition action by a higher dosage. Accordingly, the incorporation of Nitrapyrin at 2% of the applied N rate (w/w) is a recommendable dosage for local intensive greenhouse production. Further, we assessed the influence of various dosages of Nitrapyrin incorporation (0%, 0.25%, 0.5%, 2% and 5%) on the abundance and community of three groups of soil ammonia oxidizers [i.e., ammonia-oxidizing archaea (AOA), ammonia-oxidizing bacteria (AOB) and completely ammonia-oxidizing bacteria (Comammox Nitrospira)] by qPCR and high-throughput amplicon sequencing. Nitrapyrin incorporation strongly lowered both the AOB and Comammox Nitrospira abundances and their community richness even at the lowest dosage. Nitrapyrin incorporation also significantly altered the community structure of all of the tested ammonia oxidizers, and the average relative abundance of some major community members (i.e., the Nitrososphaerales Clade Nitrososphaera, Nitrososphaerales Clade A, Nitrosospira briensis Clade, Nitrosospira multiformis Clade, Comammox Nitrospira Clade A.2 and Comammox Nitrospira Clade A-associated) obviously responded to Nitrapyrin incorporation. Overall, our findings indicated that AOB and Comammox Nitrospira were more sensitive to Nitrapyrin incorporation as compared with AOA. The results obtained here highlight the importance of optimizing the type and dosage of NIs for N fertilization management in intensive greenhouse vegetable production. Nitrapyrin incorporation inhibits soil nitrification probably by suppressing the Nitrosospira multiformis Clade in the AOB community at the level tested herein.

1. Introduction

Greenhouse vegetable cultivation is one of the most important agricultural production patterns in China, since it can significantly enhance the utilization efficiency of natural resources (i.e., land, light and temperature) and thus increase the yield and farmers’ incomes in the limited cultivable land. Yet, greenhouse vegetable cultivation is commonly featured with high fertilizer input, especially nitrogen (N) fertilizer [1]. This causes a huge amount of nitrate (NO₃⁻) residual in soil, initiating a cascade of environmental issues (i.e., NO₃⁻ leaching and runoff, groundwater pollution, eutrophication and denitrification) [2].

Nitrification inhibitors (NIs) are an array of chemically synthesized or natural (biological) agents that can retard soil nitrification and, therefore, reduce the NO₃⁻ residual by synchronizing the soil N supply and crop N requirement [3,4,5], finally achieving the joint goals of increasing the N utilization efficiency and protecting the environment. Nowadays, the application of NIs has been demonstrated to be a feasible method to lower the environmental costs of intensive greenhouse vegetable production and to benefit agricultural green development [6,7].

Soil nitrification is the microbiologically driven conversion from ammonia (NH₃) to NO₃⁻, and is carried out through the following two consecutive steps: NH₃ oxidization and nitrite (NO₂⁻) oxidization. In brief, NH₃ is first oxidized to NO₂⁻ by ammonia-oxidizing archaea (AOA) and bacteria (AOB), and then the conversion from NO₂⁻ to NO₃⁻ by nitrite-oxidizing bacteria (NOB) [8]. The first step, ammonia oxidation, is the rate-limiting step of nitrification and, consequently, as the prime target for artificially interfering soil nitrification [9]. In recent years, researchers have also discovered a new group of ammonia-oxidizing bacteria (called as complete ammonia-oxidizing bacteria, Comammox). Compared with AOA and AOB, Comammox harbors an additional gene encoding nitrite oxidoreductase. This enables Comammox to combine the two steps above-mentioned into one, thereby converting NH₃ to NO₃⁻ independently [10,11]. Currently known Comammox (such as Nitrospira nitrosa, Nitrospira nitrificans, Nitrospira inopinata and Nitrospira kreftii) belong to the Nitrospira genus and are taxonomically grouped into Nitrospira lineage II [12,13,14,15]. Previously, kinetic analysis revealed an oligotrophic lifestyle by Comammox Nitrospira [16]. Yet, recently, studies have indicated their high nitrification activity and positive response to increased N input, despite the non-numerical advantages in soil [17,18,19,20]. It has been pointed out that Comammox Nitrospira in terrestrial ecosystems are abundant and active in both oligotrophic and copiotrophic environments [21]. This suggests that Comammox Nitrospira should not be ignored in exploring the microbial ecology of ammonia oxidation of high-fertility soil, such as greenhouse soil undergoing repeated fertilization and intensive cultivation. In fact, AOA, AOB and Comammox Nitrospira co-exist and work together in the soil, but their relative contributions on soil nitrification differ across habitats and change with soil environmental conditions [22,23,24,25].

At present, the most widely used NIs are Dicyandiamide (DCD), 3,4-Dimethylpyrazole phosphate (DMPP) and 2-chloro-6-(trichloromethyl) pyridine (Nitrapyrin, NP) [26,27,28]. These NIs can suppress the soil nitrification capacity by affecting the activity, quantity and community diversity and structure of ammonia oxidizers (namely, AOA, AOB and Comammox Nitrospira) [17,29,30]. Numerous studies have demonstrated the good effects of these three chemical NIs in reducing soil N loss and increasing crop N uptake [31,32,33,34]. Nevertheless, there are significant differences in the effectiveness of different types of NIs, which is closely associated with the soil properties (i.e., texture, moisture, organic matter, pH, temperature, etc.) [35,36,37]. In addition, despite AOA, AOB and Comammox Nitrospira being widely present in the terrestrial ecosystems, their biogeographical distributions differ obviously (known as niche separation) [38]. Further, the dominant/active members in soil ammonia oxidizers’ communities differ substantially across various regions, and, more importantly, these members have distinct sensitivity/resistance to NIs, leading to contrasting dosage effects [37,39,40]. Because of long-term high water and fertilizer inputs in intensive greenhouse production, as well as the relatively high soil temperature by enclosed environments, the nitrification rate in greenhouse soil is often strong, which can easily lead to fertilizer N loss and environmental pollution (such as NO₃⁻ leaching and N₂O emissions) [41,42]. Thus, exploring efficient types of NIs and their suitable dosages is of particular importance to developing a regional N fertilization management regime in intensive greenhouse production.

Xinxiang district, located in Henan province, is one of the most important greenhouse vegetable production bases in northern China. However, heavy N fertilization causes serious non-point source pollution (i.e., soil NO₃⁻ residues, groundwater pollution and gaseous N emissions) and low N utilization efficiency [43,44,45]. These restrict the green and sustainable development of the greenhouse vegetable industry in the region. We believe that the application of NIs is one of the feasible options to solve the above problems [45]. Accordingly, a soil microcosm experiment was conducted to assess the following: (1) the suitable type and dosage of NIs (DCD, DMPP and NP) in this region; (2) the effect of NI application on the abundance, community diversity and structure of soil ammonia oxidizers (AOA, AOB and Comammox Nitrospira). With the help of this experiment, we would like to answer the following two questions: (1) regarding the tested three chemical NIs, which is the more suitable for local greenhouse vegetable production; (2) how soil ammonia oxidizers responded to the NI incorporation. The results obtained from this study will help to develop a reasonable N fertilization regime in the region. Additionally, our findings can also help to advance the understanding of the active members of the soil ammonia oxidizers’ community.

2. Materials and Methods

2.1. Experimental Design and Sample Collection

The soil used for our microcosm experiment was collected from an intensively cultivated greenhouse plot located in Zhuzhuangtun village, Muye town, Xinxiang district, Henan province of central China. Muye town is one of the largest greenhouse vegetable production bases in northern China. The main crops for greenhouse cultivation are cucumber, tomato, eggplant and capsicum. Local climatic and weather conditions are as described by Liu et al. [41]. The soil type in this region is typically classified as loamy fluvo-aquic soil. After the cucumber harvest (14 July 2024), field-moist soils (0–20 cm) were collected from five randomly selected locations in the plot, and then were thoroughly mixed after removing the debris and sieved through a 2 mm mesh. The homogenized soils were immediately used for the microcosm construction. The basal soil physicochemical properties were as follows: total organic carbon, 15.26 g kg⁻¹; total nitrogen, 1.67 g kg⁻¹; NH₄⁺-N, 3.43 mg kg⁻¹; NO₃⁻-N, 178.88 mg kg⁻¹; available phosphorus, 204.67 mg kg⁻¹; available potassium, 414.68 mg kg⁻¹; a pH of 7.80 (a soil/water ratio of 1:5, w/v).

For each NI, a total of ten treatments (representing different dosages) with three replicates were set up in our microcosm experiment. The application rates for DCD were 0%, 1%, 2%, 4%, 6%, 8%, 10%, 12%, 14% and 20% of the applied N rate (w/w); for DMPP, they were 0%, 0.5%, 1%, 2%, 4%, 5%, 6%, 7%, 8% and 10% of the applied N rate (w/w); for NP, they were 0%, 0.25%, 0.5%, 1%, 2%, 2.5%, 3%, 3.5%, 4% and 5% of the applied N rate (w/w). Here, urea (46% N contained) was applied as the N source and the applied rate was 0.25 g N kg⁻¹ soil. Additionally, the control microcosms (neither urea nor NI application, with three replicates, labeled as CK) were constructed simultaneously. The DCD, NP and DMPP (with effective contents of 99%, 98% and 95%, respectively) were supplied by Shanghai Macklin Biochemical Technology Co., Ltd., Shanghai, China.

Accurately weighed quantities of urea and NIs according to each microcosm were evenly mixed, and then the mixture was mixed with 1.0 kg of fresh soil. A total of 90 mixtures (3 NIs × 10 application rates of each NI × 3 replicates) and three control soils (without urea and NI application) were transferred into independent incubation vessels (completely randomized arrangement), all of which were finally placed in an incubator at 25 °C for 60 days in the dark. The soil moisture in each microcosm was maintained at 60% of the water-holding capacity in the field throughout the experiment by weighing the incubation vessel and carefully adding sterile distilled water.

For each microcosm, soil samples were collected on day 0 (prior to incubation), and day 5, 10, 20, 30, 40, 50 and 60 from the start of the incubation. Each fresh soil sample was divided into two subsamples. One subsample was stored at −80 °C for the molecular ecological assays, and one subsample was immediately used for the measurement of the NH₄⁺-N content.

2.2. Soil Mineral N Contents

Soil NH₄⁺-N contents were extracted by a 2 mol L⁻¹ KCl solution (with a soil/water ratio of 1:5, w/v) and its contents in the filtrates were determined using an indophenol blue colorimetric assay [46].

2.3. Soil DNA Extraction and Real-Time Quantitative PCR

Here, the soil samples collected on day 20, 40 and 60 from the CK- and NP-incorporated (0%, 0.25%, 0.5%, 2% and 5%) microcosms were used for the measurement of the abundances of ammonia oxidizers by real-time quantitative PCR. The amoA gene (encoding ammonia monooxygenase subunit A) was employed as a target gene. The total microbial DNA from each soil sample was extracted using the PowerSoil DNA Isolation Kit (MoBio, San Diego, CA, USA) according to the manufacturer’s protocol. The DNA quality and purity were checked using 1.0% agarose gel electrophoresis and spectrophotometry (with the NanoDrop2000, Thermo Scientific, Waltham, MA, USA). The genomic DNA was dissolved in 50 μL of elution buffer and stored at −20 °C until further analysis.

The amoA amplicons of AOA, AOB and Comammox Nitrospira were generated from the genomic DNA. Copy numbers of the amoA gene in each sample were determined in triplicate using an ABI7300 real-time PCR system (Applied Biosystems, Foster City, CA, USA). The primer pairs and thermal cycle conditions of the PCR reactions are listed in

Table 1. PCR primers and amplification conditions used in qPCR.

Target Gene	Primer Name	Sequence	Amplification Condition	Reference
AOA amoA	Arch-amoAF Arch-amoAR	STAATGGTCTGGCTTAGACG/GCGGCCATCCATCTGTATGT	95 °C for 3 min, 40 cycles of 95 °C for 5 s, 58 °C for 30 s and 72 °C for 1 min	[47]
AOB amoA	BamoA-1F BamoA-2R	GGGGTTTCTACTGGTGGT/CCCCTCKGSAAAGCCTTCTTC		[48]
Comammox Nitrospira amoA	ComamoA-AF ComamoA-SR	AGGNGAYTGGGAYTTCTGG/CCGVACATACATRAAGCCCAT		[49]

. Each 10 μL PCR reaction mixture contained 5 μL of ChamQ SYBR Color qPCR Master Mix (2×), 0.4 μL of each primer (5 μM), 0.2 μL of ROX Reference Dye 1 (50×), 1 μL of DNA template and 3 μL of ddH₂O. The product specificities were confirmed by melting curve analyses. Amplification efficiencies of 96.79%, 103.60% and 98.78% were obtained for the archaeal, bacterial and Comammox Nitrospira amoA genes, with R² values of 0.9984, 0.9980 and 0.9982, respectively. Finally, the abundances of AOA, AOB and Comammox Nitrospira were reflected by the numbers of amoA gene copies, respectively, in 1 g of soil on a dry basis.

2.4. High-Throughput Amplicon Sequencing and Bioinformatics Analysis

Using the soil samples collected on day 20, 40 and 60 from the CK- and NP-incorporated (0%, 0.25%, 0.5%, 2% and 5%) microcosms, high-throughput amplicon sequencing was performed to assess the responses of the community diversity and structures of the soil ammonia oxidizers to various dosages of NP incorporation (Illumina MiSeq PE300 platform, San Diego, CA, USA). The primer sets and thermal cycle conditions used for amplifying the amoA gene fragments are illustrated in

Table 2. PCR primers and amplification conditions used in high-throughput amplicon sequencing.

Target Gene	Primer Name	Sequence	Amplification Condition	Reference
AOA amoA	Arch-amoAF Arch-amoAR	STAATGGTCTGGCTTAGACG/GCGGCCATCCATCTGTATGT	95 °C for 3 min, 40 cycles of 95 °C for 30 s, 60 °C for 30 s, 72 °C for 45 s and 72 °C for 10 min	[47]
AOB amoA	BamoA-1F BamoA-2R	GGGGTTTCTACTGGTGGT/CCCCTCKGSAAAGCCTTCTTC	95 °C for 3 min, 40 cycles of 95 °C for 30 s, 60 °C for 30 s, 72 °C for 45 s and 72 °C for 10 min	[48]
Comammox Nitrospira amoA	A189F C576R CA209F C576R	GGNGACTGGGAYTTYTGG/GAAGCCCATRTARTCNGCC GAYTGGAARGAYCGNCA/GAAGCCCATRTARTCNGCC	94 °C for 5 min, 20 cycles of 94 °C for 60 s, 52 °C for 50 s, 72 °C for 50 s and 72 °C for 10 min 94 °C for 5 min, 30 cycles of 94 °C for 60 s, 50 °C for 50 s, 72 °C for 50 s and 72 °C for 10 min	[50]

Comammox Nitrospira amoA gene sequences were obtained by a partial nested PCR with the primers A189F/C576R (the first round) and CA209F/C576R (the second round).

. The preliminary experiment, including six randomly selected DNA templates and four negative controls without DNA templates, was performed to optimize the amplification conditions and ensure sufficient amplification products. Finally, PCR was carried out in 20 μL reaction mixtures containing 10 μL of Pro Taq (2×), 0.8 μL of each primer (5 μM), 10 ng of DNA template and 8.4 μL of ddH₂O. Triplicate reaction mixtures for each sample were pooled together, purified using the Agarose Gel DNA Purification Kit (Axygen, Silicon Valley, CA, USA) and quantified with the NanoDrop 2000 (Thermo Scientific, Waltham, MA, USA). The PCR products were sent to Shanghai Majorbio Bio-pharm Technology Co., Ltd. (Shanghai, China) for the high-throughput sequencing. The samples were individually barcoded to enable multiplex sequencing. The barcoded PCR products from all of the samples were normalized in equimolar amounts before sequencing. The raw reads were deposited in the NCBI SRA database under BioProject accession number PRJNA1204384.

Unprocessed FASTQ files were extracted for the downstream analysis. The overlapping paired-end reads were assembled using FLASH software (version 1.2.11). This step was not carried out for the archaeal amoA gene reads because they did not overlap under the paired-end sequencing, and, therefore, only the forward reads were used for the evaluation of the community structure of AOA [51]. Further, quality filtration and data analysis of the sequences were performed using Mothur software (version 1.30.2), as described by Lu et al. [52]. In total, we obtained 2,450,242, 2,180,374 and 2,016,898 high-quality sequences of archaeal, bacterial and Comammox Nitrospira amoA genes, respectively, from the 54 soil samples tested. The average length of these high-quality sequences was 274 bp for the archaeal amoA gene, 451 bp for the bacterial amoA gene and 365 bp for the Comammox Nitrospira amoA gene, respectively. Based on a sequence similarity of 97%, the sequences were divided into different operational taxonomic units (OTUs) by Uparse software (version 11), and the most abundant sequence of each OTU was used as the representative sequence. Neighbor-joining phylogenetic trees of the representative sequences were constructed through MEGA software (version 6.0) with 1000 bootstrap replicates. Reference sequences and the nomenclatures for various archaeal and bacterial amoA gene clusters were from previous reports by Alves et al. [53] and Zhang et al. [54], and the various Comammox Nitrospira amoA gene clusters were from a previous study by Zou et al. [55]. Before the community assessment, all of the samples were rarefied randomly to an identical number of reads according to the sample with the lowest reads. The rarefaction curve indicated that the numbers of reads sampled were enough to reflect the actual states of the soil microbiome.

2.5. Statistical Analysis

Statistical analysis was carried out by a one-way ANOVA (Duncan’s multiple comparisons at a 95% level of probability) to determine the significant differences between the treatments using SPSS software for Windows (version 25.0). The soil ammonia oxidizers’ abundances were Log₁₀-transformed. The mean values and standard error (n = 3) are also reported in this paper. Additionally, a two-way ANOVA was applied to explore the interactive effect of the treatment (T) and incubation time (S). The occurrence of the relationships between various data was determined by one-variable regression modeling. The alpha diversity of the soil microbial community was estimated using Mothur software (version 1.30.2). The principal coordinate analysis (PCoA), based on an algorithm for Bray–Curtis dissimilarity, was conducted to assess the changes in the community structures of the soil ammonia oxidizers after the NI application. An analysis of similarities (ANOSIM) was performed to examine whether there were significant differences in the community structures between the treatments. Meanwhile, a two-way permutational multivariate analysis of variance (PERMANOVA) was also used to assess the separate and interactive effects of the treatment (T) and incubation time (S) on the microbial community structure. The procedures and methodology details of the analyses above-mentioned are as described by Ren et al. [56].

3. Results

3.1. The Change in Soil NH₄⁺-N Content

After the DCD incorporation, the soil NH₄⁺-N content rapidly increased, and the peak occurred on day 5; thereafter, the soil NH₄⁺-N content decreased until day 20; after day 20, there was no significant difference in the soil NH₄⁺-N content between the treatments, except for the soil samples collected on day 40 (Figure 1A). The response of the soil NH₄⁺-N content to the DMPP incorporation was completely similar with that to the DCD incorporation (Figure 1B). These results suggested that there was a transient inhibitory effect of DCD and DMPP on soil nitrification (less than five days), even at their highest dosages.

As illustrated in Figure 1C, in all of the NP-incorporated microcosms, the soil NH₄⁺-N content rapidly increased from day 0 to day 5 and then kept stable from day 5 to day 10. After day 10, the soil NH₄⁺-N content in the NP-0.25% and NP-0.5% microcosms gradually decreased, and, on day 40, the soil NH₄⁺-N content in these microcosms had no obvious difference with the NP-free microcosms. Regarding the microcosms with the NP-1% and NP-2% incorporations, the soil NH₄⁺-N content decreased since day 20 and day 30, respectively. For the microcosms with the NP-2.5%, NP-3%, NP-3.5% and NP-4% incorporations, the soil NH₄⁺-N contents all decreased from day 40. The soil NH₄⁺-N content in the NP-5% microcosms had a significant decline from day 50. In general, a greater dose of NP incorporation inhibited the soil nitrification with more time.

3.2. The Change in the Abundance of Soil Ammonia Oxidizers After NP Incorporation

Given that there was a persistent inhibition effect of the NP incorporation on soil nitrification, we further assessed the response of the abundances of soil ammonia oxidizers to various dosages of NP incorporation (NP-0%, NP-0.25%, NP-0.5%, NP-2% and NP-5%) through using the soil samples collected on days 20, 40 and 60. Across all of the soil samples tested here, AOA outnumbered AOB and Comammox Nitrospira (Figure 2A–C). In brief, the AOA abundance ranged from 4.37 × 10⁷ to 1.10 × 10⁹ amoA gene copies g⁻¹ (2.53 × 10⁸ amoA gene copies g⁻¹ on average, with a coefficient of variation of 65.73%); the AOB abundance ranged from 4.74 × 10⁶ to 1.22 × 10⁸ amoA gene copies g⁻¹ (3.00 × 10⁷ amoA gene copies g⁻¹ on average, with a coefficient of variation of 93.64%); Comammox Nitrospira ranged from 7.24 × 10⁴ to 3.27 × 10⁶ amoA gene copies g⁻¹ (8.20 × 10⁵ amoA gene copies g⁻¹ on average, with a coefficient of variation of 103.42%).

Both the treatment and incubation time significantly affected the soil AOA abundance (Figure 2A). At day 20, compared with NP-0%, the AOA abundances in NP-2% and NP-5% significantly decreased by 41.18% and 61.30%, respectively; at day 40, only in NP-5% did the soil AOA abundance show a significant decline by 69.01% relative to NP-0%. These results suggest that the AOA abundance was suppressed only by a high dosage of NP incorporation. Even so, such suppression was also time-dependent, and finally disappeared at day 60.

For AOB, its abundance was significantly inhibited by NP incorporation at each of the incubation times (Figure 2B). In detail, compared to NP-0%, the AOB abundances in NP-0.25%, NP-0.5%, NP-2% and NP-5% significantly decreased by 72.09%, 68.75%, 78.22% and 82.56% at day 20, respectively; significantly reduced by 91.20%, 87.47%, 85.26% and 92.09% at day 40, respectively; significantly lowered by 86.63%, 83.64%, 91.28% and 90.71% at day 60, respectively.

In line with AOB, the abundance of Comammox Nitrospira was also significantly suppressed by the NP incorporation at each of the incubation times (Figure 2C). Specifically, in comparison with NP-0%, the Comammox Nitrospira abundances in NP-0.25%, NP-0.5%, NP-2% and NP-5% significantly decreased by 61.15%, 71.13%, 83.62% and 89.32% at day 20, respectively; significantly reduced by 87.42%, 72.79%, 68.54% and 65.77% at day 40, respectively; significantly lowered by 78.02%, 78.60%, 83.25% and 85.22% at day 60, respectively.

A low dosage of the NP incorporation, even at NP-0.25%, significantly inhibited the abundances of both AOB and Comammox Nitrospira, suggesting their higher sensitivity to the NP incorporation relative to AOA. Compared with CK (neither urea nor NP incorporation), the AOB abundance in NP-0% significantly increased at each of the incubation times, indicating a positive effect of the urea addition on the AOB abundance. Nevertheless, such a positive effect was not observed for the abundance of AOA or Comammox Nitrospira.

In addition, there was a significantly negative relationship in the soil NH₄⁺-N content and AOA abundance (Figure 2D), and the AOB abundance (Figure 2E), but not for the Comammox Nitrospira abundance (Figure 2F).

3.3. The Effect of NP Incorporation on AOA Community

Neither the treatment nor the incubation time altered the AOA community richness, which was fully demonstrated by the number of observed species and Chao1 index (Figure 3A,B). Nevertheless, both the treatment and incubation time significantly altered the AOA community diversity (Figure 3C). At day 20, there was no significant difference in the Shannon index between the treatments; at day 40, the Shannon index in NP-0.25% and NP-2% significantly decreased by 5.54% and 19.50% as compared to NP-0%, respectively, but that in NP-5% significantly increased by 6.27%; at day 60, the Shannon index in NP-2% and NP-5% significantly decreased 12.19% and 11.99% relative to NP-0%, respectively. Additionally, at each of the incubation times, the Shannon index of the AOA community in NP-0% did not differ from that in CK, indicating that the urea amendment had no effect on the AOA community diversity. The PCoA analysis and ANOSIM testing demonstrated that there was a significant difference in the AOA community structure among the tested soil samples (Figure 3D). Furthermore, the two-way PERMANOVA testing confirmed that the AOA community structure was affected by the treatment and incubation time (p < 0.05), with a more prominent effect by the former. However, regarding the AOA community richness, diversity and structure, each of them did not link to the soil NH₄⁺-N content (p > 0.05) (Figure 4).

A total of 115 OTUs for AOA was obtained in this study, of which there were 16 OTUs in the average relative abundance exceeding 0.1%. These 16 OTUs occupied near 99% of the total recovered sequences. The phylogenetic analysis showed that these dominant OTUs belonged to the Nitrososphaerales Clade Nitrososphaera, Nitrososphaerales Clade A and Ca. Nitrosotaleales Clade Nitrosotalea (Figure S1), with the first two completely dominating the community (Figure 3E). Also, both the treatment and incubation time significantly altered the average relative abundances of these two predominant community members. For the Nitrososphaerales Clade Nitrososphaera, at day 20, its average relative abundance in NP-0.25% and NP-0.5% significantly increased by 98.65% and 81.79% compared to NP-0%, respectively; at day 40, its average relative abundance increased first and then decreased along with the dosage of the NP incorporation; at day 60, its average relative abundance generally increased along with the dosage of the NP incorporation (Figure S2A). For Nitrososphaerales Clade A, at day 20, its average relative abundance in NP-0.25% and NP-0.5% significantly decreased by 64.44% and 57.89% compared to NP-0%, respectively; at day 40, its average relative abundance decreased first and then increased along with the dosage of the NP incorporation; at day 60, its average relative abundance generally decreased along with the dosage of the NP incorporation (Figure S2B). In addition, neither the Nitrososphaerales Clade Nitrososphaera nor Nitrososphaerales Clade A correlated to the soil NH₄⁺-N content in terms of the average relative abundance (p > 0.05) (Figure S2C,D).

3.4. The Effect of NP Incorporation on AOB Community

The NP incorporation significantly inhibited the AOB community richness. For the number of observed species (Figure 5A) and Chao1 index (Figure 5B), which, in the treatments with the NP incorporation, were significantly lower than those in the treatments without the NP incorporation. In addition, the NP incorporation also altered the AOB community diversity. At day 20, the Shannon index in NP-2% was significantly lowered by 14.93% compared to that in NP-0% (Figure 5C). The NP incorporation affected the AOB community richness more strongly than the community diversity. There was a significant difference in the AOB community structure among the soil samples tested here, and, further, the treatment exerted a stronger influence on the community structure than did the incubation time (Figure 5D). The AOB community richness, diversity and structure were significantly correlated to the soil NH₄⁺-N content (p < 0.05) (Figure 4).

For AOB, a total of 1163 OTUs were obtained, of which there were 72 OTUs with average relative abundances of more than 0.1%. These 72 OTUs altogether occupied 91% of the recovered sequences. The phylogenetic analysis indicated that these dominant OTUs were affiliated into two bacterial genera, namely, Nitrosospira and Nitrosomonas (Figure S3). Nitrosospira completely dominated the AOB community (Figure 5E) and was subdivided into the Nitrosospira briensis Clade, Nitrosospira multiformis Clade, Nitrosospira sp. Nsp65/L115 Clade and Nitrosospira sp. Np39-19 Clade (Figure S3). The Nitrosospira briensis Clade and Nitrosospira multiformis Clade were the major members in the AOB community across all of the soil samples (Figure 5E). Both the treatment and incubation time significantly changed the average relative abundance of the Nitrosospira briensis Clade; however, the average relative abundance of the Nitrosospira multiformis Clade was influenced by the treatment rather than the incubation time (Figure 5E). For the Nitrosospira briensis Clade, when compared to NP-0%, at day 20, its average relative abundance in NP-0.25%, NP-0.5%, NP-2% and NP-5% significantly increased by 94.30%, 60.70%, 114.63% and 115.15%, respectively; at day 40, it significantly increased by 121.14%, 129.51%, 76.04% and 132.82%, respectively; at day 60, it significantly increased by 126.38%, 121.60%, 123.77% and 138.17%, respectively (Figure S4A). For the Nitrosospira multiformis Clade, its average relative abundance was almost completely suppressed by the NP incorporation at each of the incubation times (in average, it decreased by 17.71-fold, 12.77-fold and 10.60-fold on day 20, day 40 and day 60, respectively) (Figure S4B). Compared to CK, the average relative abundance of the Nitrosospira briensis Clade in NP-0% decreased, but the opposite was true for that of the Nitrosospira multiformis Clade, suggesting their contrasting response to the urea amendment. Moreover, the average relative abundance of Nitrosospira briensis Clade was positively correlated to the soil NH₄⁺-N content (p < 0.0001) (Figure S4C); however, the average relative abundance of the Nitrosospira multiformis Clade was negatively associated with the soil NH₄⁺-N content (p < 0.0001) (Figure S4D).

3.5. The Effect of NP Incorporation on Comammox Nitrospira Community

The NP incorporation significantly inhibited the Comammox Nitrospira community richness. The number of observed species (Figure 6A) and Chao1 index (Figure 6B) in the treatments with the NP incorporation were significantly lower than those in the treatments without the NP incorporation. Also, the NP incorporation suppressed the Comammox Nitrospira community diversity. At day 20, the Shannon index in NP-0.5%, NP-2% and NP-5% were significantly lowered by 36.90%, 61.29% and 47.39% as compared to NP-0%, respectively (Figure 6C). There was no significant difference in the number of observed species, or Chao1 and Shannon indices between NP-0% and CK, indicating that the urea amendment did not affect the community richness and diversity of Comammox Nitrospira. Although there was a significant difference in the Comammox Nitrospira community structure between the soil samples tested here, such a difference was attributable to the treatment rather than the incubation time (Figure 6D). The Comammox Nitrospira community richness, diversity and structure were significantly correlated to the soil NH₄⁺-N content (p < 0.05) (Figure 4).

For Comammox Nitrospira, there were 12 OTUs with average relative abundances exceeding 0.1%, and these dominant OTUs altogether accounted for more than 98% of the total recovered sequences. The phylogenetic analysis revealed that 11 out of the 12 OTUs were affiliated within Comammox Nitrospira Clade A, and the remaining OTU belonged to Comammox Nitrospira Clade B (Figure S5). Further, the OTUs belonging to Clade A were subdivided into Clade A.2 and Clade A-associated (Figure S5). Comammox Nitrospira Clade A.2 almost completely dominated the community (Figure 6E). Both for Comammox Nitrospira Clade A.2 and Clade A-associated, their average relative abundances were influenced by the treatment more strongly as compared to the incubation time (Figure 6E). For the Comammox Nitrospira Clade A.2, at day 60, its average relative abundance in NP-5% significantly decreased by 6.30% relative to NP-0% (Figure S6A). In terms of Comammox Nitrospira Clade A-associated, at day 60, its average relative abundance in NP-5% significantly increased by 11.50-fold relative to NP-0% (Figure S6B). Generally, the urea amendment (i.e., NP-0% vs. CK) increased the average relative abundance of Comammox Nitrospira Clade A.2 but decreased that of Comammox Nitrospira Clade A-associated. Yet, the average relative abundances of both Comammox Nitrospira Clade A.2 and Clade A-associated did not link to the soil NH₄⁺-N content (Figure S6C,D).

4. Discussion

4.1. NP as a Suitable NI in the Intensively Cultivated Greenhouse Soil Tested Here

As commonly used commercial chemical NIs worldwide, DCD, DMPP and NP have been confirmed as having a positive effect on reducing N losses (mainly through NO₃⁻ leaching and N₂O emissions), improving the N use efficiency and increasing crop N uptake [57,58,59]. Yet, the efficiency of these three NIs to reduce N losses differs from each other, which is directly attributable to the duration of the inhibition action and the applied dosages under different soil environmental conditions or soil types [59,60,61,62]. In soil, urea is enzymatically degraded into NH₄⁺-N first and then NH₄⁺-N is rapidly converted into NO₃⁻-N. Given that NO₃⁻-N has a high mobility and is the substrate for denitrification, decelerating the conversion of NH₄⁺-N to NO₃⁻-N is crucial for reducing the soil NO₃⁻-N residual, and the associated N losses, while increasing the crop N uptake. Hence, clarifying the suitable type and dosage of NIs is responsible for developing the best crop N management practice at a regional scale. In this study, DCD and DMPP had a transient inhibitory influence on soil nitrification because the soil NH₄⁺-N content in the DCD/DMPP-incorporated microcosms declined from day 5, regardless of the dosages (Figure 1A,B). By contrast, the NP-incorporated microcosms held a high soil NH₄⁺-N content over a longer time (Figure 1C), suggesting an effective and persistent suppression on soil nitrification. For this reason, NP could be considered as a promising soil nitrification inhibitor for intensive regional greenhouse production. In accordance with the present study, Fan et al. [63], Xi et al. [64] and Zhang et al. [65] also reported an effective inhibition to nitrification by NP in intensively managed greenhouse vegetable-cropped soil.

Herein, there were several possible explanations for the weak inhibition capacity on soil nitrification by DCD and DMPP. First, there was a significantly negative linear relationship between the soil NH₄⁺-N content and both the AOA and AOB abundances (Figure 2D,E), suggesting that these two groups of ammonia oxidizers actively drove soil nitrification. However, previous researchers have pointed out that DMPP and DCD mainly target AOB, but not AOA, whereas NP can inhibit the activities of both AOA and AOB [66,67]. This, to a certain extent, better explains the inhibition effectiveness of NP on soil nitrification as compared to DCD or DMPP. There are distinct modes of action for NIs to inhibit nitrification. Specifically, DCD and DMPP act as copper-chelating agents acting on the B subunit of ammonia monooxygenase (AMO), consequently inactivating the functional group of AMO; NP suppresses the activity of ammonia oxidizers by blocking the AMO enzymatic pathway (altering AMO’s structure and impairing its reaction with substrate) [36,68,69]. Interestingly, the B subunit of AMO for AOB contains an active site that catalyzes reactions using copper, while the active site for AOA is seated at the C or X subunit of AMO [36,70]. Second, there was high soil organic matter content in the soil tested here. NIs can be adsorbed onto soil particles with a high level of organic matter, resulting in a low availability in the soil and exerting less impact on ammonia oxidizers [63,71,72,73]. By comparison with NP, DCD and DMPP may be more easily adsorbed because of their lower molecular weights. On the other hand, DMPP is positively charged and can be adsorbed by negatively charged soil particles, and negatively charged DCD becomes sorbed to metal oxides in a more alkaline soil [73]. Third, concerning the soil physicochemical properties, pH is thought to be one of the most important factors influencing NI efficiency because it can affect the mobility and degradation of NIs in soils [74]. The tested soil in this study had an alkaline pH value (pH = 7.8). Previous studies demonstrated that the inhibition efficacy of DCD and especially DMPP decreased in calcareous soil (pH > 7.0) [73]; conversely, NP had a better efficacy in inhibiting the nitrification rate and reducing N₂O emissions in alkaline soils (pH = 7.6 and 8.15) [69,75]. Finally, the solubility of NP is much less than that of DCD and DMPP, thus reducing the spatial separation of NP from NH₄⁺ [69]. Notably, these possible reasons require careful verification in future work.

There was an obvious dosage effect of the NP incorporation on the inhibition of soil nitrification (Figure 1C). A higher dosage of the NP incorporation contributed to a longer inhibition time. In our study region, the farmers’ N fertilization in greenhouse vegetable production is commonly carried out in split, with 40% of N fertilizer as a basal application at transplanting and the remaining as topdressing. Starting from one month after transplanting, the topdressing is applied in combination with irrigation at an interval of 7–10 days. Therefore, from transplanting to the first instance of topdressing, this time segment easily contributes to soil N residues and losses due to the relatively low N demand by the plants. This means that NIs must exert a persistent inhibition on soil nitrification for at least one month. According to the results obtained from this study, the incorporation of NP at a dosage of 2% of the applied N rate inhibited soil nitrification over one month (Figure 1C). Hence, we believe that this is a recommendable dosage to be applied to local intensive greenhouse production. On the other hand, we must admit that such a dosage obtained here still needs to be verified in the field condition, because some of the soil biotic (i.e., microbial activity) and abiotic (i.e., moisture, temperature and nutrient availability) properties in the microcosm experiments were greatly different from actual agricultural production (which are more complex and ever-changing). Overall, our study highlights the importance of optimizing the NI type and dosage in regional N fertilization management.

4.2. NP Incorporation Suppressed AOB and Comammox Nitrospira Abundances Strongly

In the current study, under the condition of high dosages of the NP incorporation, the AOA abundance significantly declined at days 20 and 40, but significantly increased at day 60 (Figure 2A). This suggests that the inhibition of the NP incorporation on AOA growth is reversible/recoverable, particularly at a high dosage level. The underlying mechanism is that a high dosage of the NP incorporation strongly restrains the AOB and Comammox Nitrospira abundances (Figure 2B,C), thereby releasing a large living space for AOA [76,77,78]. Through searching the Web of Science Core Collection database (with “nitrapyrin” as a topic), we summarized the existing observations by previous researchers regarding the influence of NP incorporation on the abundances of soil ammonia oxidizers (

Table 3. Summary on the influence of the NP incorporation on the abundance of soil ammonia oxidizers observed from previous studies using a qPCR assay.

Experimental Type	Soil Type	Dosage Applied	Change in Microbial Abundance			Reference
			AOA	AOB	COM
Laboratory incubation	Calcareous soil	1%	Not sig.	Declined	—	[69]
Laboratory incubation	Black soil	0.1%	Increased	Declined	Increased	[36]
Laboratory incubation	Red soil	0.1%	Increased	Declined	—	[36]
Laboratory incubation	Purple soil	0.1%	Increased	Declined	—	[36]
Field experiment	Yellow clay soil	Not mentioned	Decreased	Not sig.	—	[82]
Laboratory incubation	Pasture soil	Not mentioned	Not sig.	Declined	Declined	[30]
Laboratory incubation	Arable soil	Not mentioned	Not sig.	Declined	Declined	[30]
Laboratory incubation	Sandy loam soil	At an equivalent rate of 2.5 L hm−2	Not sig.	Declined	—	[83]
Field experiment	Fimi-Orthic Anthrosols	Not mentioned	Decreased	Declined	—	[62]
Field experiment	Oxisols	0.25%	Decreased	Not sig.	—	[84]
Laboratory incubation	Haplic Arenosol	9 µg active ingredient g−1 of dry soil	—	Not sig.	—	[85]
Laboratory incubation	Vertosol	At an equivalent rate of 2.5 L hm−2	Increased	Declined	—	[76]
Laboratory incubation	Tenosol	At an equivalent rate of 2.5 L hm−2	Increased	Declined	—	[76]
Laboratory incubation	Sodosol	At an equivalent rate of 2.5 L hm−2	Increased	Declined	—	[76]
Laboratory incubation	Calcarosol	At an equivalent rate of 2.5 L hm−2	Increased	Not sig.	—	[76]
Field experiment	Paddy soil	0.2%	Declined	Not sig.	—	[86]
Laboratory incubation	Loamy soil	50 mg kg−1 of dry soil	Not sig.	Declined	Not sig.	[17]
Laboratory incubation	Red soil	0.1%	Not sig.	Declined	—	[64]
Laboratory incubation	Black soil	0.2%	—	Declined	—	[87]
Laboratory incubation	Cambisol—acidic	0.86 mg kg−1 of dry soil	Declined	Not sig.	—	[77]
Laboratory incubation	Cambisol—acidic	5 mg kg−1 of dry soil	Declined	Declined	—	[77]
Laboratory incubation	Cambisol—alkaline	0.86 mg kg−1 of dry soil	Not sig.	Declined	—	[77]
Laboratory incubation	Cambisol—alkaline	5 mg kg−1 of dry soil	Declined	Declined	—	[77]
Laboratory incubation	Sandy soil—Tenosol	5 mg kg−1 of dry soil	Not sig.	Declined	—	[88]
Laboratory incubation	Sandy soil—Hydrosol	5 mg kg−1 of dry soil	—	Declined	—	[88]
Laboratory incubation	Upland alluvial soil	0.3 mg kg−1 of dry soil	Declined	Declined	—	[89]
Laboratory incubation	Paddy soil	0.3 mg kg−1 of dry soil	Declined	Declined	—	[89]
Laboratory incubation	Upland black soil	0.3 mg kg−1 of dry soil	Not sig.	Not sig.	—	[89]

Not sig. means no significant change in the AOA, AOB or Comammox Nitrospira abundance after NP incorporation. The symbol “—” means that the abundance of this group of ammonia oxidizers was not tested. COM, Comammox Nitrospira.

). In line with the results obtained here, some of the researchers reported significantly increased AOA abundances after the NP incorporation (Table 3). A previous meta-analysis indicated that the NP incorporation slightly affected and even increased the soil AOA abundance [79]. Also, in terms of the major AOA community members, the average relative abundance of the Nitrososphaerales Clade Nitrososphaera rather than Nitrososphaerales Clade A significantly increased at day 60 in the microcosms receiving high dosages of the NP incorporation (NP-2% and NP-5%) (Figure 3E and Figure S2A,B), suggesting that the increased AOA abundance observed here was associated with the rapid growth of the Nitrososphaerales Clade Nitrososphaera, and also suggesting that this AOA Clade might be more resilient/competitive relative to other Clades [80,81].

Unlike AOA, the NP incorporation strongly suppressed both the AOB and Comammox Nitrospira abundances throughout the experiment, regardless of the dosages (Figure 2B,C). Most of the previous studies demonstrated the significant suppression of the NP incorporation on the AOB abundance across the experiment types, soil types or applied dosages (Table 3), powerfully supporting our observation. However, to date, only very few studies have evaluated the influence of NP incorporation on soil Comammox Nitrospira abundance; meanwhile, the studies involved were conducted in the condition of laboratory incubation (Table 3). Particularly, there were completely distinct observations (not significant, significantly increased or significantly declined) on the response of the Comammox Nitrospira abundance to the NP incorporation by previous researchers (Table 3). The disparity is probably associated with the soil backgrounds and the dosages applied. As a newly discovered functional group of soil ammonia oxidization, Comammox Nitrospira, the response of its abundance to NP incorporation urgently needs to be assessed in diverse soil types and dosages, and in both field and laboratory conditions. Moreover, Comammox Nitrospira did not correlate to the soil NH₄⁺-N content (Figure 2F), probably suggesting that the population size of Comammox Nitrospira has a minor effect on soil nitrification at the level tested herein. Yet, this inference needs to be carefully evaluated in future work.

4.3. NP Incorporation Significantly Affected the Communities of Soil Ammonia Oxidizers

With the help of high-throughput amplicon sequencing, we also examined the influence of the NP incorporation on the community richness, diversity and structure of the soil ammonia oxidizers. The NP incorporation affected the community richness and diversity of AOB and Comammox Nitrospira more strongly as compared with those of AOA (Figure 3A–C, Figure 5A–C and Figure 6A–C). Particularly, the NP incorporation violently suppressed the community richness of AOB and Comammox Nitrospira throughout the experiment, even at the condition of a low dosage. These results indicate the higher sensitivity of the AOB and Comammox Nitrospira communities to the NP incorporation relative to AOA at the overall community level. Meanwhile, there were significantly negative correlations of the soil NH₄⁺-N content with both the AOB and Comammox Nitrospira richness (Figure 4), suggesting that the NP incorporation inhibited soil nitrification probably via suppressing or even killing some specific taxa in their communities. As illustrated in

Table 4. Summary on the effect of the NP incorporation on the community richness, diversity and structure of soil ammonia oxidizers observed by previous researchers using high-throughput amplicon sequencing with the amoA gene as a target.

Experimental Type	Soil Type	Dosage Applied	Main Findings			Reference
			AOA	AOB	COM
Laboratory incubation	Calcareous soil	1%	Richness: not sig. Diversity: not sig. Structure: not sig.	Richness: not sig. Diversity: not sig. Structure: significantly changed; average relative abundance of the Nitrosospira sp. Nsp17 Clade declined, whereas that of the Nitrosospira sp. 9SS1 and Nitrosospira briensis Clades increased	—	[69]
Field experiment	Oxisols	0.25%	Richness: not mentioned Diversity: not mentioned Structure: not sig.	Richness: not mentioned Diversity: not mentioned Structure: not sig.	—	[84]
Laboratory incubation	Loamy soil	50 mg kg−1 of dry soil	Richness: not sig. Diversity: not sig. Structure: not sig.	Richness: not sig. Diversity: not sig. Structure: significantly changed; reduced the average relative abundance of the Nitrosospira sp. NI5 Clade	Richness: declined Diversity: declined Structure: significantly changed; slightly increased the average relative abundance of Comammox Nitrospira Clade A without statistical significance	[17]
Laboratory incubation	Cambisol—Acidic	0.86 mg kg−1 of dry soil	Richness: not sig. Diversity: not sig. Structure: not sig.	Richness: not sig. Diversity: not sig. Structure: significantly changed; increased the average relative abundance of some ASVs belonging to the Nitrosospira sp. Nsp5 Clade	—	[77]
Laboratory incubation	Cambisol—Acidic	5 mg kg−1 of dry soil	Richness: not sig. Diversity: not sig. Structure: significantly changed; decreased the average relative abundance of an unclassified ASV	Richness: not sig. Diversity: not sig. Structure: significantly changed; lowered the average relative abundance of some ASVs belonging to the Nitrosospira briensis and Nitrosospira sp. Nsp65 Clades	—	[77]
Laboratory incubation	Cambisol—Alkaline	0.86 mg kg−1 of dry soil	Richness: not sig. Diversity: not sig. Structure: not sig.	Richness: not sig. Diversity: declined Structure: significantly changed; the majority of the ASVs were significantly affected by the NP belonging to the Nitrosospira briensis group	—	[77]
Laboratory incubation	Cambisol—Alkaline	5 mg kg−1 of dry soil	Richness: not sig. Diversity: not sig. Structure: significantly changed; suppressed the average relative abundance of the ASVs belonging to the Nitrososphaerales ε-2.2 Clade, but favored the ASVs belonging to the Nitrososphaerales γ Clade	Richness: declined Diversity: declined Structure: significantly changed; the majority of the ASVs were significantly affected by the NP belonging to the Nitrosospira briensis group	—	[77]

Not sig. means no significant change. The symbol “—” means that this group of ammonia oxidizers was not tested. COM, Comammox Nitrospira. Community richness is reflected by the number of observed species or the Chao1 index, and the community diversity is reflected by the Shannon index.

, based on the Web of Science Core Collection database (using the “nitrapyrin” as a topic), we also summarized the existing observations by previous studies regarding the effect of NP incorporation on the community richness, diversity and structure of soil ammonia oxidizers. We indeed observed that the AOA community richness and diversity commonly did not change after the NP incorporation, and, on the contrary, the response of the AOB community richness and diversity to the NP incorporation relied on the soil type and dosage (Table 4). Notably, to date, only one study has reported on the influence of NP incorporation on the Comammox Nitrospira community (as shown in Table 4), and clearly revealed the declined community richness and diversity of Comammox Nitrospira, supporting the observations obtained here.

For all three of the groups of soil ammonia oxidizers tested here, the NP incorporation significantly changed their community structures (Figure 3D, Figure 5D and Figure 6D). Correspondingly, such alterations could be confirmed by the significant changes in the average relative abundance of the major community members (Figure 3E, Figure 5E and Figure 6E). For the AOA community, in CK, soil exceeding 70% of archaeal amoA sequences fell within Nitrososphaerales Clade A (Figure 3E and Figure S2B), suggesting that this Clade held a dominant position in our tested greenhouse soil. Liu et al. [41] also reported that Nitrososphaerales Clade A was not only prevalent but also predominant in intensively greenhouse vegetable-cultivated soil. Another study pointed out that Nitrososphaerales Clade A had a poor NH₃-oxidizing capacity by determining the in situ and potential gross nitrification rates in the soil dominated by this AOA Clade [53]. In terms of the Nitrososphaerales Clade Nitrososphaera, Gao et al. [90] reported that it was functionally active in the acidic paddy soil by applying ¹⁵N-DNA stable isotope probing; nevertheless, other researchers revealed that this AOA Clade favored a neutral soil pH and was inhibited by heavy N fertilization [91,92]. In a representative strain in the Nitrososphaerales Clade Nitrososphaera, Nitrososphaera viennensis EN76, its growth and ammonia-oxidizing activity were inhibited by the NP addition during the microbial plate culture [93]; moreover, this strain was also significantly inhibited by biological NIs, such as caffeic acid, phenylalanine, vanillic acid and vanillin [94]. Yet, Yin et al. [81] confirmed that, when AOBs were blocked by DMPP or 1-octyne across three arable soils (Gleyic Stagnic Anthrosol, Luvic Phaeozem and Aquic Inceptisol), the recovery of AOA communities was related to the increased abundance of Nitrososphaera viennesis. Despite the Nitrososphaerales Clade Nitrososphaera (also referred as Thaumarchaeota group 1.1b) having a potential adaption to copiotrophic environments, its growth remains susceptible to competition with AOB [95,96,97]. Herein, the NP incorporation generally decreased the average relative abundance of Nitrososphaerales Clade A but increased that of the Nitrososphaerales Clade Nitrososphaera (Figure S2A,B). In the NP-incorporated soil, the increased average relative abundance of the Nitrososphaerales Clade Nitrososphaera was likely highly correlated to the weakened competition with AOB (because a lot of the AOB was inhibited by the NP incorporation). In other words, such a weak competition stimulated/helped the growth of the Nitrososphaerales Clade Nitrososphaera.

For the AOB community, the NP incorporation increased the average relative abundance of the Nitrosospira briensis Clade but strongly inhibited that of the Nitrosospira multiformis Clade (Figure 5E and Figure S4A,B). Previous studies in calcareous soil [69] and Cambisol [77] also found an increased average relative abundance of the Nitrosospira briensis Clade after the NP incorporation. It is notable that a recently published study indicated that the response of the Nitrosospira briensis Clade to NP incorporation depended on the soil pH [77]. The Nitrosospira multiformis Clade was previously referred to as AOB Nitrosospira cluster 3a.2. This Clade had been identified as a key nitrifier group involved in N₂O emissions in fertilized agricultural soil [98]. Further, it is reported that NP incorporation inhibited AOB growth, with a prominent inhibition to the abundance and activity of AOB Nitrosospira cluster 3a.2, and thus reduced soil cumulative N₂O emissions [98]. With the help of the plate culture method, Papadopoulou et al. [67] and Shen et al. [93] indicated the strong inhibition of the NP addition on the activity and growth of the Nitrosospira multiformis strain. These results support the results obtained in this study. More importantly, of the major soil ammonia oxidizers’ community members detected here (i.e., the Nitrososphaerales Clade Nitrososphaera, Nitrososphaerales Clade A, Nitrosospira briensis Clade, Nitrosospira multiformis Clade, Comammox Nitrospira Clade A.2 and Comammox Nitrospira Clade A-associated), only the average relative abundance of the Nitrosospira multiformis Clade was significantly and negatively correlated with the soil NH₄⁺-N content, suggesting that this Clade might be an active taxon driving soil nitrification, and the NP incorporation suppressed the soil nitrification by suppressing its growth. However, this inference needs to be carefully verified by applying the DNA stable isotope probing method in the future.

Regarding Comammox Nitrospira, the Clade A almost dominated the whole community, and, at the finer taxonomic level, more than 90% of the total recovered sequences fell within the Clade A.2 (Figure 6E and Figure S6A). These results also agree with some previous reports that most of Comammox Nitrospira in agricultural soils belong to Clade A.2 [99,100,101]. A previous study pointed out that Comammox Nitrospira Clade A can adapt to copiotrophic environments and positively respond to N input/fertilization [21], which explains the universality and dominant position of this Clade in most agricultural soils. Interestingly, compared to CK, in NP-0% microcosms, the average relative abundance of Comammox Nitrospira Clade A-associated significantly decreased (Figure S6B), suggesting that not all taxa affiliated with Clade A harbor the copiotrophic lifestyle and can adapt to N-amended soils, implying that the ecophysiology and environmental adaptability of various subgroups of Comammox Nitrospira are more complex than previously thought. Wen et al. [102] found that DMPP incorporation decreased the Comammox Nitrospira community diversity but increased the community richness in tea (Camellia sinensis) plantation soils. Shah et al. [103] confirmed that DMPP incorporation did not alter the diversity of the Comammox Nitrospira community, while it significantly changed the community structure in a grassland soil located on the South Island of New Zealand. Hsu et al. [29] reported that DCD incorporation altered the Comammox Nitrospira community structure and inhibited Clade B in Templeton sandy loam soil. Yet, to the best of our knowledge, the observations involved in how the Comammox Nitrospira community structure responds to NP incorporation remain rare so far (Table 4). The only report demonstrated that the average relative abundance of both Clade A and B did not significantly change after the NP incorporation in a loamy soil [17]. Similarly, in this study, the average relative abundance of both Comammox Nitrospira Clade A.2 and Clade A-associated were insensitive to the NP incorporation, with statistically significant alterations only occurring on day 60 under the highest dosage applied (Figure S6A,B). Also, the average relative abundance of these two community members of Comammox Nitrospira did not link to the soil NH₄⁺-N content, probably suggesting their minor role in soil nitrification. Further exploration targeting the contribution of the overall Comammox Nitrospira community and various community members to soil nitrification under the condition of NP incorporation is extremely necessary.

5. Conclusions

According to the results obtained from our microcosm experiments, Nitrapyrin is a considerably recommendable nitrification inhibitor in the intensive greenhouse vegetable production system because of its longer duration of inhibition action relative to DCD and DMPP. The NP incorporation provided a more profound influence on AOB and Comammox Nitrospira as compared to AOA. For AOB and Comammox Nitrospira, Nitrapyrin incorporation powerfully inhibited their abundances and community richness, and significantly altered their community structure. The Nitrosospira multiformis Clade in the AOB community may be an active nitrifier group in the intensively cultivated greenhouse soil tested here, and may be sensitive/fragile to Nitrapyrin incorporation.