Effects of Soil pH on Gaseous Nitrogen Loss Pathway via Feammox Process

: The application of N fertilizer is one of the most critical soil acidiﬁcation factors in China, and soil acidiﬁcation signiﬁcantly alters biogeochemical processes such as N loss. Anaerobic ammonium oxidation coupled with iron reduction (Feammox) is an important biological process for N loss in natural environments, with the end-products of N 2 , NO 2 − and NO 3 − . However, the response of Feammox pathways to soil pH ﬂuctuation has not been thoroughly studied. In the current study, Feammox pathways and microbial communities were explored through a slurry culture experiment with an artiﬁcially adjusted pH combined with a 15 N isotope tracing technique and molecular biotechnology. Results showed signiﬁcant differences in the gaseous N loss through Feammox (0.42–0.97 mg N kg − 1 d − 1 ) under different pH conditions. The gaseous N loss pathways were signiﬁcantly affected by the pH, and Feammox to N 2 was the predominant pathway in low-pH incubations. The proportion of N loss caused by Feammox coupled with denitriﬁcation increased as the soil pH increased. The gaseous N loss through Feammox increased by 43.9% when the soil pH decreased from 6.5 to 5.0. Fe-reducing bacteria, such as Ochrobactrum , Sphingomonas , and Clostridium increased signiﬁcantly in lower pH incubations. Overall, this study demonstrated the effects of soil pH on Feammox pathways and extended the understanding of the N biogeochemical cycle in acidic soil.


Introduction
With the decrease of arable land area and the decrease of soil fertility, the dependence of agriculture on chemical fertilizers is increasingly serious [1]. The overuse of Nitrogen (N) fertilizer has caused critical environmental problems, such as soil acidification [2]. It was found that the soil pH in the major grain-producing regions in China has decreased significantly since 1980, and the contribution of fertilization to H + is much higher than that of acid deposition [3]. Soil acidification has adverse effects on terrestrial ecosystems and causes a series of secondary problems, such as the deterioration of soil properties and the activation of heavy metals [4].
A large amount of N fertilizer is applied to cultivated land in the agricultural production process. However, only about 35% of the N can be used by crops, and the remaining N enters the water and atmosphere through leaching and microbial metabolism [5]. Recently, anaerobic ammonium oxidation coupled with iron reduction (Feammox) has been reported as a novel microbial process involved in the soil N cycle [6][7][8][9]. NH 4 + is directly oxidized to N 2 (Equation (1)), NO 2 − (Equation (2)), and NO 3 − (Equation (3)), by microorganisms using Fe(III) as an electron acceptor under anaerobic conditions in the Feammox process [10,11]. Compared to the nitrification-denitrification process, the Feammox process shortens the N Sustainability 2021, 13, 10393 2 of 11 cycle time, and makes a great contribution to N loss in ecosystems. It was estimated that a loss of 7.8-61.0 kg N ha −1 y −1 was associated with Feammox, representing 3.9-31.0% of the N fertilizer applied to paddy soils in southern China [7]. 3Fe(OH) 3 + 5H + +NH 4 + → 3Fe 2+ + 9H 2 O + 0.5N 2 ∆ r G m = −245 kJ mol −1 (1) 6Fe(OH) 3 + 10H + +NH 4 + → 6Fe 2+ + 16H 2 O + NO 2 − ∆ r G m = −164 kJ mol −1 (2) 8Fe(OH) 3 + 14H + +NH 4 + → 8Fe 2+ + 21H 2 O + NO 3 − ∆ r G m = −207 kJ mol −1 (3) As an important environmental factor, the soil pH significantly influences the microbial N cycle, including the processes of aerobic ammonia oxidation, denitrification, nitrification, and anammox. Denitrification and anammox processes are generally inhibited under acidic conditions, whereas Feammox might favor lower-pH environments as the process needs a ready supply of H + ions [12,13]. A canonical correlation analysis showed that pH is one of the key factors affecting the rate of Feammox according to the analysis of 66 soil samples collected from different soil/sediments [14]. Moreover, Acidimicrobiaceae A6 has been identified as the only Feammox functional microorganism which can oxidize NH 4 + to NO 2 − , was also isolated from acidic wetland sediment and showed the highest NH 4 + oxidation rates at pH 4 [11]. Nevertheless, there are three end-products (N 2 , NO 2 − and NO 3 − ) in Feammox process (Equations (1)-(3)), and the effect of soil pH changes on Feammox pathways has not been thoroughly studied. According to thermodynamic calculations, Feammox to N 2 is more likely to occur than the other two pathways (Equation (1)). Although the above conclusion is supported by recent studies, two other Feammox pathways with the end-products of NO 2 − and NO 3 − are noteworthy [6,7]. Especially in acidic soils/sediments, the activities of microorganisms such as denitrification and anammox are inhibited, and NO 2 − produced through Feammox is prone to accumulation, which can further lead to water pollution or greenhouse gas (e.g., N 2 O) emission [12]. With the burning of fossil fuels and the extensive use of chemical fertilizers, the situation of soil acidification of cultivated land is gradually increasing [3]. Therefore, it is of great significance to explore the influence of soil pH changes on the rates and pathways of Feammox for controlling N loss in agricultural fertilization and reducing the risk of environmental pollution.
Red soil is one of the most important soil types in South China, and its high content of iron oxides lays a foundation for the occurrence of Feammox. Meanwhile, soil acidification is common in southern China, and it restricts crop growth and poses a severe threat to food security [3]. The main objectives of this study were to explore the effects of soil pH on the pathways of Feammox and the subsequent N transformation. The microbial communities of Feammox and their responses to soil pH also have been investigated.

Isotopic Tracer Incubation
Soil samples were collected from Yiliang County, Kunming, China (24 • 47 N, 103 • 1 E) in November 2017. The average annual temperature of the site is 16.3 • C, and the annual precipitation is 1000 mm. The region has a subtropical monsoon and the soil type is red soil, with long-term crop cultivation (mainly dryland crops). Five samples were collected within a range of 10 × 10 m in the area (0-10 cm depth) and then mixed together. The samples were transported to the laboratory in an icebox within 24 h and divided into 3 subsamples. One subsample was air-dried for analyzing the soil properties, the second subsample was used for isotope incubation, and the remaining subsample was stored in the freezer at −80 • C. The soil pH was 6.5 ± 0.3, and the water content was 48.8 ± 0.2%. The total Fe, TOC, NH 4 + -N, and NO 3 − -N contents were 87.7 ± 3.5 g·kg −1 , 70.0 ± 0.8 g·kg −1 , 40.9 ± 0.3 mg·kg −1 , and 1.9 ± 0.2 mg·kg −1 , respectively. NO 2 − -N was not detectable in the soil.
Isotope tracer incubation was conducted in an anaerobic glove box (Coy Laboratory Products, Grass Lake, MI, USA) filled with ultra-pure He using the procedure slightly modified from Ding et al. [7]. Prior to incubation, sterile anoxic deionized water was added to the Sustainability 2021, 13, 10393 3 of 11 soils at a ratio of 3:1 (v/w) and pre-incubated in an anaerobic glove box in the dark at 20 • C for 1 week to remove background O 2 and NO x − . After pre-incubation, the O 2 and NO x − concentration was maintained at a low level (<0.1 mg·L −1 and <1 mg·kg −1 , respectively). After pre-incubation, 21.5 g of homogenized slurries were transferred into 50 mL serum vials. The soil acidification process was simulated by manually adjusting the slurry pH. Three pH groups (5.0 ± 0.1, 6.5 ± 0.1, and 8.0 ± 0.1) were set in the incubation, representing acidic, weakly acidic, and alkaline soil pH. 1 mM NaOH and HCl solutions were used to adjust the pH of the slurries to the preset values. The pH of the slurries was adjusted again to the set values after 12 h to reduce the influence of the soil acid-base buffer on the experiment. In each pH group, 3 treatments (n = 5 per treatment) were established to detect Feammox, namely (1) the control with only sterile anoxic deionized water; (2) 15 ). It has been reported that C 2 H 2 (acetylene) can inhibit the reduction of N 2 O to N 2 in denitrification and the activity of anammox bacteria [15,16]. The serum vials were then sealed with butyl rubber stoppers, and the headspace was exchanged with ultra-pure Ar for 5 min. The 15 NH 4 Cl solution was prepared using sterile anoxic deionized water, and 0.5 mL of 15 NH 4 Cl solution was injected into the serum vials with sterile syringes so that the final concentration of 15 NH 4 Cl was 5 mM. For the C 2 H 2 treatment, 7.5 mL of headspace gas in each vial was removed and replaced with 7.5 mL of C 2 H 2 purified by H 2 SO 4 to reach 30% (v/v) C 2 H 2 in the headspace [7]. Isotope tracer incubations were operated at 20 • C in a constant temperature incubator, and all the vials were shaken vigorously to homogenize the treatment solutions daily. A set of vials was destructively sampled in an anaerobic glove box at 0, 10, 20, and 30 d. All the vials were shaken vigorously before sampling to equilibrate the gas between the dissolved and gaseous phases. Five milliliters of the headspace gas were extracted with sterile syringes and injected into a 12 mL pre-vacuumed glass tube (Labco Co., Ltd., High Wycombe, UK). The 15 N enrichment in N 2 was determined by isotope ratio mass spectrometry (Thermo Fisher Scientific, Waltham, MA, USA). The concentrations of 15 N 2 were calculated according to the production of N 2 and its concentrations above the natural abundance [7].
The potential Feammox rate was conservatively estimated based on the difference in the 30 N 2 production between the treatment with 15 NH 4 Cl and the control. The N 2 production rates were calculated from the changes in the N 2 concentrations between two given times. The Feammox pathways were distinguished by adding C 2 H 2 , and the concentration of C 2 H 2 used in this study was able to completely inhibit the denitrification and anammox processes [7]. The Feammox-N 2 rate (Equation (1)) was estimated by the 30 N 2 production rate in the 15 NH 4 + + C 2 H 2 treatment [6].

Anaerobic Chemodenitrification Experiment
In order to explore the effects of pH on the chemodenitrification reaction under anaerobic conditions, a series of experiments was conducted. The Fe 2+ and NO 2 − solutions were prepared with anoxic deionized water. A 10 mL 5 mM Fe 2+ solution was injected into 50 mL serum vials, and then the initial pH was adjusted to 4.0-7.0 using NaOH and HCl solutions. After exchanging the headspace with ultra-pure Ar, the serum vials were sealed with butyl rubber stoppers. An NO 2 − solution was added into the serum vials with sterile syringes to make the concentration of NO 2 − in the reaction system reach 2 mM. The serum vials with only NO 2 − or Fe 2+ were set as controls at pH 5.0. The experiments were repeated in triplicate. The serum vials were incubated at 30 • C in a constant temperature incubator in the dark for 100 h. The supernatant was extracted at setting time, and the concentration of the NO 2 − was determined by ion chromatography (Thermo Fisher Scientific, Waltham, MA, USA).

Physicochemical Analysis
The soil pH was determined using a pH meter (Mettler-Toledo, Zurich, Switzerland), and the ratio of deionized water to soil was 2.5:1 (v/w). The total organic C (TOC) content was determined by the potassium dichromate oxidation method [17]. The NH 4 + and NO x − were extracted with a 2 M KCl solution for 1 h and determined by a continuous-flow nutrient auto-analyzer (Skalar Analytic, Breda, The Netherlands). The total extractable Fe(III) and Fe(II) in the soil were determined according to the slightly modified method of Ding et al. [7]. Briefly, 1.0 g of soil was extracted with 5 mL of 0.50 M HCl and 5 mL of 0.25 M hydroxylamine hydrochloride in 0.25 M HCl for 2 h at 30 • C in an anaerobic glove box. The ferrozine method was used to measure the total Fe and Fe(II) contents [18]. The 15 N enrichment in the N 2 was determined by isotope ratio mass spectrometry (Thermo Fisher Scientific, Waltham, MA, USA).

Microbial Community Analysis
DNA samples were extracted from 1.0 g of soil slurry collected after 30 d of incubation and amplified using the primer pair 515F and 806R for the V4 region of the 16S rRNA gene as described by the manufacture's protocol. Sequencing was conducted at a commercial company (Majorbio, Shanghai, China). The microbial community functional profiles were predicted using PICRUSt (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States). DNA samples were extracted from 1.0 g of soil slurry collected after 30 days' incubation using a Fast DNA SPIN Kit for soil (MP Biomedical, Solon, OH, USA), according to the manufacture's protocol. The quality and purity of the extracted DNA were determined using the NanoDrop-2000 spectrophotometer (NanoDrop Technologies, Wilmington, NC, USA) and then stored in the refrigerator at -20 • C. The extracted DNA was amplified using the primer pair 515F and 806R for V4 region of the 16S rRNA gene. PCRs were performed in a total volume of 30 µL containing 15 µL. The microbial community functional profiles were predicted by using PICRUSt (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States). The 16S rRNA gene sequences were compared in the KEGG database (www.genome.jp/kegg/ pathway.html, (accessed on 7 June 2018)) to analyze the relative abundance of the KEGG orthologous groups associated with the N cycle in metabolic pathways. The predicted functional data were clustered according to the abundance distribution of functional groups or the similarity between samples. The functional groups and samples were sorted according to the clustering results and then presented by the heat map. R-project was used for cluster analysis and heat map drawing (www.R-project.org, (accessed on 20 June 2018)).

Statistical Analyses
One-way analysis of variance and Pearson's correlation analysis were conducted using SPSS 22.0 (SPSS Inc., Chicago, IL, USA), and p < 0.05 was set to denote the significance level. One-way ANOVA analyses were used to calculate the variance between obtained data under different pH values. Pearson's correlation analysis was performed to evaluate the relationships among the nitrite concentrations, Feammox rates, and Fe(III) reduction rates. The figures in this study were drawn using Origin 9.0 (Origin Lab, Northampton, MA, USA).

pH-Dependent Feammox Rates and Pathways
For all the pH groups, a significant amount of 30  30 N 2 production rates were significantly (p < 0.05) higher in the 5.0 pH group than those in the 6.5 and 8.0 pH groups, varying from 0.94 ± 0.08 mg N kg −1 d −1 in the 5.0 pH group to 0.42 ± 0.03 mg N kg −1 d −1 in the 15 NH 4 + addition treatments in the 8.0 pH group. The variations in the 29 N 2 production rate with pH were identical to those of the 30 N 2 . However, the 29 N 2 production rates were higher than those of the 30 N 2 in all the groups.
rates. The figures in this study were drawn using Origin 9.0 (Origin Lab, Northampton, MA, USA).

pH-Dependent Feammox Rates and Pathways
For all the pH groups, a significant amount of 30 N2 was detected in the 15 NH4 + and 15 NH4 + + C2H2 treatments, whereas no 30 N2 accumulated in the control (Figure 1a). The 30 N2 production rates were significantly (p < 0.05) higher in the 5.0 pH group than those in the 6.5 and 8.0 pH groups, varying from 0.94 ± 0.08 mg N kg −1 d −1 in the 5.0 pH group to 0.42 ± 0.03 mg N kg −1 d −1 in the 15 NH4 + addition treatments in the 8.0 pH group. The variations in the 29 N2 production rate with pH were identical to those of the 30 N2. However, the 29 N2 production rates were higher than those of the 30 N2 in all the groups. In this study, the soil slurries were strictly pre-incubated in an anoxic environment to remove indigenous molecular O2, NO3 − , and NO2 − . Strict anaerobic conditions were maintained during the 15 N isotope incubation. Meanwhile, the 16S rRNA sequencing results showed that no typical anammox bacteria were found in any of the experimental treatments. Therefore, the aerobic nitrification and anammox processes can be ruled out. Under these conditions, the Feammox to N2 (Equation (1)) and Feammox-generated 15 NOx − (Equations (2) and (3)) reduction by denitrification are the only potential biological pathways for 30 N2 production. Moreover, 15 NH4 + promoted the Fe(III) reduction rates in the incubations, thereby further providing robust evidence for the occurrence of the Feammox process (Supplementary Figure S1). The Feammox rates in our incubation (0.42-0.94 mg N kg −1 d −1 ) were of the same order of magnitude as those reported in natural environments ( Figure 1 and Supplementary Table S1). The production rates of 30 N2 and 29 N2 in the 5.0 pH groups (with and without C2H2) were significantly (p < 0.05) higher than those of the other pH groups, and there were higher initial pH values with lower Feammox rates ( Figure 1). The significantly higher Feammox rates in the acid red soil than those in natural environments indicated that acid soil is more beneficial to the occurrence of Feammox (Supplementary, Table S1). However, it has been reported that a higher pH and lower Eh are beneficial to Feammox in mangrove wetlands [19], and the Feammox process also occurs in alkaline sediments (e.g., the intertidal wetlands of the Yangtze Estuary) [8]. The responses of the Feammox to soil pH fluctuations are not thoroughly understood. Fe oxides in the soil will release more soluble Fe(III) when the pH decreases [20], and there In this study, the soil slurries were strictly pre-incubated in an anoxic environment to remove indigenous molecular O 2 , NO 3 − , and NO 2 − . Strict anaerobic conditions were maintained during the 15 N isotope incubation. Meanwhile, the 16S rRNA sequencing results showed that no typical anammox bacteria were found in any of the experimental treatments. Therefore, the aerobic nitrification and anammox processes can be ruled out. Under these conditions, the Feammox to N 2 (Equation (1)) and Feammox-generated 15 (2) and (3)) reduction by denitrification are the only potential biological pathways for 30 N 2 production. Moreover, 15 NH 4 + promoted the Fe(III) reduction rates in the incubations, thereby further providing robust evidence for the occurrence of the Feammox process (Supplementary Figure S1). The Feammox rates in our incubation (0.42-0.94 mg N kg −1 d −1 ) were of the same order of magnitude as those reported in natural environments (Figure 1 and Supplementary Table S1). The production rates of 30 N 2 and 29 N 2 in the 5.0 pH groups (with and without C 2 H 2 ) were significantly (p < 0.05) higher than those of the other pH groups, and there were higher initial pH values with lower Feammox rates (Figure 1). The significantly higher Feammox rates in the acid red soil than those in natural environments indicated that acid soil is more beneficial to the occurrence of Feammox (Supplementary, Table S1). However, it has been reported that a higher pH and lower Eh are beneficial to Feammox in mangrove wetlands [19], and the Feammox process also occurs in alkaline sediments (e.g., the intertidal wetlands of the Yangtze Estuary) [8]. The responses of the Feammox to soil pH fluctuations are not thoroughly understood. Fe oxides in the soil will release more soluble Fe(III) when the pH decreases [20], and there might be higher microbially reducible Fe(III) contents in acidic soil, which can be beneficial for providing more electron acceptors for the Feammox.
The thermodynamic calculations showed that the Feammox-N 2 pathway could be conducted under a wide range of pH conditions, but the Feammox-NO 2 − pathway occurs only when the pH is less than 6.5 [6]. Previous studies have shown that Feammox-N 2 is the main pathway of gaseous N loss in the Feammox process (>50%), but the effect of soil pH on the Feammox pathways has been rarely reported. The Feammox-N 2 pathway accounted for 47.4-86.2% of the total 30 N 2 loss in all the pH groups, and the proportion was significantly (p < 0.05) higher in the low-pH group (Supplementary Figure S2). The gaseous N loss ratio contributed by the Feammox coupled with denitrification (Feammox-denitrification) was 52.6% at a pH of 8.0. Although the PICRUSt results showed a significant increase in the denitrifying microbial abundance at a low pH, the gaseous N loss rate from the Feammoxdenitrification pathway remained low in the pH 5.0 group (Supplementary Figure S2). This may be due to the large amount of NO 2 − produced by Feammox under low pH conditions, which promoted the growth of denitrifying bacteria. However, the optimal pH value of the denitrifying bacteria was 7.0-9.0, and the activity of the denitrifying bacteria decreased under low pH conditions [12].

Potential Transformation Pathways of NO 2 −
The concentrations of NO 3 − -N and NO 2 − -N in the serum samples were measured at the end of the incubation (Figure 2). The concentrations of NO 2 − -N were significantly (p < 0.05) higher than that of the NO 3 − -N in each pH group. For the three different pH groups, the concentrations of NO 2 − -N in the 15 NH 4 + and 15 NH 4 + + C 2 H 2 treatments at a pH of 5.0 were 4.84 ± 0.46 mg·kg −1 and 6.13 ± 0.47 mg·kg −1 , respectively, which were significantly higher than those in the other pH conditions (p < 0.05). The linear regression analysis showed a positive correlation between the NO 2 − -N contents and the Feammox rate of 30 N 2 in all the treatments (Supplementary Figure S3), which further indicated that NO 2 − is an important intermediate product of the Feammox-N 2 pathway (Equation (2)). NO 2 − is one of the primary end-products of Feammox in natural and wastewater treatment systems and plays a vital role in the transformation of the Fe cycle [10,21,22]. The Feammox-NO 2 − pathway and the further transformation process of NO 2 − are worthy of attention.
The thermodynamic calculations showed that the Feammox-N2 pathway could be conducted under a wide range of pH conditions, but the Feammox-NO2 − pathway occurs only when the pH is less than 6.5 [6]. Previous studies have shown that Feammox-N2 is the main pathway of gaseous N loss in the Feammox process (>50%), but the effect of soil pH on the Feammox pathways has been rarely reported. The Feammox-N2 pathway accounted for 47.4-86.2% of the total 30 N2 loss in all the pH groups, and the proportion was significantly (p < 0.05) higher in the low-pH group (Supplementary Figure S2). The gaseous N loss ratio contributed by the Feammox coupled with denitrification (Feammox-denitrification) was 52.6% at a pH of 8.0. Although the PICRUSt results showed a significant increase in the denitrifying microbial abundance at a low pH, the gaseous N loss rate from the Feammox-denitrification pathway remained low in the pH 5.0 group (Supplementary Figure S2). This may be due to the large amount of NO2 − produced by Feammox under low pH conditions, which promoted the growth of denitrifying bacteria. However, the optimal pH value of the denitrifying bacteria was 7.0-9.0, and the activity of the denitrifying bacteria decreased under low pH conditions [12].

Potential Transformation Pathways of NO2 −
The concentrations of NO3 − -N and NO2 − -N in the serum samples were measured at the end of the incubation (Figure 2). The concentrations of NO2 − -N were significantly (p < 0.05) higher than that of the NO3 − -N in each pH group. For the three different pH groups, the concentrations of NO2 − -N in the 15 NH4 + and 15 NH4 + + C2H2 treatments at a pH of 5.0 were 4.84 ± 0.46 mg·kg −1 and 6.13 ± 0.47 mg·kg −1 , respectively, which were significantly higher than those in the other pH conditions (p < 0.05). The linear regression analysis showed a positive correlation between the NO2 − -N contents and the Feammox rate of 30 N2 in all the treatments (Supplementary Figure S3), which further indicated that NO2 − is an important intermediate product of the Feammox-N2 pathway (Equation (2)). NO2 − is one of the primary end-products of Feammox in natural and wastewater treatment systems and plays a vital role in the transformation of the Fe cycle [10,21,22]. The Feammox-NO2 − pathway and the further transformation process of NO2 − are worthy of attention. Previous studies have shown that the Fe(III) reduction rates are significantly positively correlated with the N2 production rates [6,7,9]. The fluctuation in the Fe(III) reduction rate was not clear among the three pH groups in our study (Supplementary Figure S1). However, the 30 N2 production rate was significantly increased in the 5.0 pH groups. The gaseous N production rate in the 15 NH4 + + C2H2 treatment was also significantly increased at a pH of 5.0. This meant that the chemodenitrification reaction occurred and played an important role in the 5.0 pH groups because the biological reduction of NOx − was completely inhibited. Previous studies have shown that the Fe(III) reduction rates are significantly positively correlated with the N 2 production rates [6,7,9]. The fluctuation in the Fe(III) reduction rate was not clear among the three pH groups in our study (Supplementary Figure S1). However, the 30 N 2 production rate was significantly increased in the 5.0 pH groups. The gaseous N production rate in the 15 NH 4 + + C 2 H 2 treatment was also significantly increased at a pH of 5.0. This meant that the chemodenitrification reaction occurred and played an important role in the 5.0 pH groups because the biological reduction of NO x − was completely inhibited. It has been reported that NO 2 − can be reduced to N 2 with Fe(II) through chemodenitrification under anaerobic conditions (Equation (4)), which may be one of the potential pathways of N 2 loss in the Feammox process [21,23]. The kinetics of the chemodenitrification reaction under different pH values indicated that the NO 2 − reduction rates increased with the decrease in the pH (Figure 3). It has been demonstrated that chemodenitrification makes a greater contribution to NO 2 − decomposition in sterilized acid soil [24]. The increase in the NO 2 − reduction rate at a low pH may be due to the higher proportion of HNO 2 in the liquid phase. The protonation of NO 2 − promotes the fracture of N-O, thereby resulting in a more robust oxidation of HNO 2 than NO 2 − [25].
This confirmed that Feammox-generated Fe(II) was more likely to react with NO 2 − in the low-pH groups and further promote the N and Fe biogeochemical cycles.
under anaerobic conditions (Equation (4)), which may be one of the potential pathways of N2 loss in the Feammox process [21,23]. The kinetics of the chemodenitrification reaction under different pH values indicated that the NO2 − reduction rates increased with the decrease in the pH (Figure 3). It has been demonstrated that chemodenitrification makes a greater contribution to NO2 − decomposition in sterilized acid soil [24]. The increase in the NO2 − reduction rate at a low pH may be due to the higher proportion of HNO2 in the liquid phase. The protonation of NO2 − promotes the fracture of N-O, thereby resulting in a more robust oxidation of HNO2 than NO2 − [25]. This confirmed that Feammox-generated Fe(II) was more likely to react with NO2 − in the low-pH groups and further promote the N and Fe biogeochemical cycles.
4Fe(II) + 2NO2 − +8H + → 4Fe(III) + N2 + 9H2O (4) Amorphous Fe oxides in the soil will release more soluble Fe(III) when the pH decreases [20], which would promote the Feammox process and provide more Fe(II). Studies have shown that the surface of soil minerals can reaggregate the Fe(II) produced by microorganisms and accelerate the transfer of electrons to NO2 − . Meanwhile, the Fe(II) adsorbed on the soil minerals has a higher oxidizability in the reaction between Fe(II) and NO2 − compared with that of dissolved Fe 2+ [26]. Therefore, soil acidification will facilitate Fe(II) generation in the Feammox process and the chemical reduction of NO2 − induced by Fe(II). The Fe(III) reduction was initiated by the Feammox process, and the generated Fe(II) was oxidized to Fe(III) by the NO2 − -N reduction. Owing to the lack of continuous data on the NO2 − -N concentration, the extent of Feammox-NO2 − could not be accurately calculated in our study. Amorphous Fe oxides in the soil will release more soluble Fe(III) when the pH decreases [20], which would promote the Feammox process and provide more Fe(II). Studies have shown that the surface of soil minerals can reaggregate the Fe(II) produced by microorganisms and accelerate the transfer of electrons to NO 2 − . Meanwhile, the Fe(II) adsorbed on the soil minerals has a higher oxidizability in the reaction between Fe(II) and NO 2 − compared with that of dissolved Fe 2+ [26]. Therefore, soil acidification will facilitate Fe(II) generation in the Feammox process and the chemical reduction of NO 2 − induced by Fe(II). The Fe(III) reduction was initiated by the Feammox process, and the generated Fe(II) was oxidized to Fe(III) by the NO 2 − -N reduction. Owing to the lack of continuous data on the NO 2 − -N concentration, the extent of Feammox-NO 2 − could not be accurately calculated in our study.

Evolution of the Microbial Community in Feammox
Proteobacteria was the dominant phylum in the 5.0 pH group in this study, with a relative abundance of 43.3-44.9% (Figure 4). It has been reported that Ochrobactrum and Sphingomonas from Proteobacteria might contribute to a metals and NO x − reduction, and are more abundant in acidic incubations [27][28][29][30]. The relative abundance of Firmicutes significantly increased (p < 0.05) at a pH of 5.0. Clostridium accounted for 6.68-15.59% of all the microorganisms, which was possibly because the genus is more suitable to more acidic environments. Clostridium species have been reported to reduce Fe(III) and NO 3 − using acetate as an electron donor [31,32]. Our results also demonstrated that the Acidimicrobiaceae family had a higher abundance in the low-pH group. Although the genus information could not be obtained at the 97% OTU similarity level for Acidimicrobiaceae in this study, Acidimicrobiaceae has been classified as Fe-reducing and Feammox bacteria [22]. Acidimicrobiaceae bacterium A6 has been widely reported and identified as the functional microorganism responsible for the Feammox process [14,22].
Sphingomonas from Proteobacteria might contribute to a metals and NOx reduction, and are more abundant in acidic incubations [27][28][29][30]. The relative abundance of Firmicutes significantly increased (p < 0.05) at a pH of 5.0. Clostridium accounted for 6.68-15.59% of all the microorganisms, which was possibly because the genus is more suitable to more acidic environments. Clostridium species have been reported to reduce Fe(III) and NO3 − using acetate as an electron donor [31,32]. Our results also demonstrated that the Acidimicrobiaceae family had a higher abundance in the low-pH group. Although the genus information could not be obtained at the 97% OTU similarity level for Acidimicrobiaceae in this study, Acidimicrobiaceae has been classified as Fe-reducing and Feammox bacteria [22]. Acidimicrobiaceae bacterium A6 has been widely reported and identified as the functional microorganism responsible for the Feammox process [14,22].  Owing to the positive relationship between Fe-reducing bacteria and Feammox rates, researchers believe that Fe-reducing bacteria play a vital role in Feammox. The reported Fe-reducing bacteria have shown diverse communities in ecosystems, including Geobacter, Shewanella, Anaeromyxobacter, and Pseudomonas [9, 33,34]. However, the Feammox-N 2 functional microorganisms have not yet been identified. Geobacter and Shewanella were not detected in any of the treatments in this study. Geobacter would disappear during long-term Feammox culture [22]. In this study, the diversity of Fe-reducing bacteria changed significantly with the soil pH. Ochrobactrum, Sphingomonas, and Clostridium may play significant roles in Fe reduction in the Feammox culture system at a pH of 5.0.
The effects of pH on the microbial community functional profiles were investigated using PICRUSt ( Figure 5). The results showed that the relative abundances of the clusters of orthologous groups in the 5.0 pH groups, which were associated with NO 3 − and NO 2 − reduction or transformation (K00370-K00374, K00376, and K02305), were significantly higher than those at other pH values (p < 0.05). In this study, the detected NO x − in the incubation was generated from the Feammox process. The low environmental pH promoted the Feammox-NO 2 − pathway (Equation (2)) and stimulated denitrifying microbial growth, which was consistent with the PICRUSt results. Although most denitrifiers have been reported in neutral and weakly alkaline environments, acidic environments can stimulate the growth of specific denitrifiers, such as Ochrobactrum and Clostridium [28,32]. In addition, the long-term culture of the Feammox system also showed that the abundance of the nirS and nirK genes increased, thereby demonstrating that denitrification was also active during the Feammox process [14,22].
icantly with the soil pH. Ochrobactrum, Sphingomonas, and Clostridium may play significant roles in Fe reduction in the Feammox culture system at a pH of 5.0.
The effects of pH on the microbial community functional profiles were investigated using PICRUSt ( Figure 5). The results showed that the relative abundances of the clusters of orthologous groups in the 5.0 pH groups, which were associated with NO3 − and NO2 − reduction or transformation (K00370-K00374, K00376, and K02305), were significantly higher than those at other pH values (p < 0.05). In this study, the detected NOx − in the incubation was generated from the Feammox process. The low environmental pH promoted the Feammox-NO2 − pathway (Equation (2)) and stimulated denitrifying microbial growth, which was consistent with the PICRUSt results. Although most denitrifiers have been reported in neutral and weakly alkaline environments, acidic environments can stimulate the growth of specific denitrifiers, such as Ochrobactrum and Clostridium [28,32]. In addition, the long-term culture of the Feammox system also showed that the abundance of the nirS and nirK genes increased, thereby demonstrating that denitrification was also active during the Feammox process [14,22].

Environmental Implications
Compared with that of previous studies, the high Feammox rate could be attributed to the low pH and high Fe content in the soil explored in this study. The soil was collected from the subtropical region of Southern China, which has a severe soil acidification problem. Based on the rates obtained from isotope incubations, the gaseous N loss via the Feammox process increased by 43.9% when the pH decreased from 6.5 to 5.0 in red soils. This result indicated that Feammox could be a vital pathway for N loss in acidic soils. However, the strict anaerobic conditions and high NH 4 + concentrations in the incubations possibly led to the overestimation of the Feammox rates, and precise evaluation is needed to perform in situ studies.
Theoretically, the oxidation of 1 mol of NH 4 + will consume 5-14 protons in the Feammox process, and the efficiency of neutralizing H + is much higher than that of lime, which was generally used to mitigate soil acidification (Equation (1)). However, the excessive lime application reduces the availability of nutrients in the soil, and additional agricultural measures are required to thoroughly mix the lime with the soil [35,36]. It has been reported that the pH value increases with the incubation and NH 4 + dosage in the Feammox process [10]. Similarly, the pH increased significantly at the end of our incubation in the 5.0 pH group (data not shown). Therefore, the Feammox process may play an essential role in the acid-base equilibrium in soils. Feammox is theoretically feasible for remediating acidified soil, and its effects still need to be investigated through further experiments.

Conclusions
This study demonstrated that Feammox is a crucial gaseous N loss pathway in acidic upland soil (0.42-0.97 mg N kg −1 d −1 ), and the Feammox pathways were also affected by the soil pH. Specifically, N 2 was mainly directly generated by the Feammox process in acidic soils, while most of the N 2 loss was attributed to Feammox-coupled denitrification in high-pH soils. The gaseous N loss through Feammox increased by 43.9% when the soil pH decreased from 6.5 to 5.0. In addition, NO 2 − is an important end-product of Feammox in acidic soil and is possibly reduced by Fe(II) through the abiotic/biotic process. Moreover, the relative abundances of Acidimicrobiaceae and Fe-reducing bacteria increased significantly under acidic conditions. This study provides experimental evidence of the influence of soil acidification on the Feammox pathways and new insights into the interaction between N application and soil acidification.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/su131810393/s1, Figure S1: Fe(III) reduction rates measured in the isotope tracer incubations, Figure S2: The contribution of Feammox to N 2 pathway (Feammox-N 2 ) to gaseous N loss, Figure S3: Pearson's correlations of nitrite concentrations with both Feammox rates (a) and Fe(III) reduction rates (b), Table S1: Research on Feammox in natural environments.  Data Availability Statement: The data are not publicly available due to privacy issue.