Contribution of Anammox to Nitrogen Removal in Two Temperate Forest Soils

ABSTRACT

Anaerobic ammonium oxidation with nitrite reduction to dinitrogen (termed anammox) has been reported to be an important process for removing fixed nitrogen (N) in marine ecosystems and in some agricultural and wetland soils. However, its importance in upland forest soils has never been quantified. In this study, we evaluated the occurrence of anammox activity in two temperate forest soils collected from northeastern China. With ¹⁵N-labeled NO₃⁻ incubation, we found that the combined potential of the N₂ production rates of anammox and codenitrification ranged from 0.01 ± 0.01 to 1.2 ± 0.18 nmol N per gram of soil per hour, contributing 0.5% to 14.4% of the total N₂ production along the soil profile. Denitrification was the main pathway of N₂ production and accounted for 85.6% to 99.5% of the total N₂ production. Further labeling experiments with ¹⁵NH₄⁺ and ¹⁵NO₂⁻ indicated that codenitrification was present in the mixed forest soil. Codenitrification and anammox accounted for 2% to 12% and 1% to 7% of the total N₂ production, respectively. Two anammox species, “Candidatus Brocadia fulgida” and “Candidatus Jettenia asiatica,” were detected in this study but in very low abundance (as indicated by the hzsB gene). Our results demonstrated that the anammox process occurs in forest soils, but the contribution to N₂ loss might be low in these ecosystems. More research is necessary to determine the activities of different N₂ releasing pathways in different forest soils.

IMPORTANCE In this study, we examined the anammox activity in temperate upland forest soils using the ¹⁵N isotope technique. We found that the anammox process contributed little to the N₂ production rate in the studied forest soil. Two anammox organisms, “Candidatus Brocadia fulgida” and “Candidatus Jettenia asiatica,” were detected. In addition, we found that codenitrification was another N₂ production pathway in forest soils. Our results could contribute to the understanding of soil gaseous N losses and microbial controls in forest soils.

INTRODUCTION

Traditionally, denitrification is considered the only pathway for active nitrogen (N) removal in the form of dinitrogen (N₂) and is an important mechanism for explaining the N process in both terrestrial and aquatic ecosystems (1). However, the discovery of anaerobic ammonium oxidation (anammox) with nitrite reduction to N₂ in the early 1990s challenged these views (2). Since then, the anammox process has been reported to occur in marine ecosystems, such as sediments (3, 4), the anoxic water column (5), and oxygen minimum zones (OMZs) (6–9). Subsequently, the anammox process was discovered in estuary sediments (10–14), river sediments (15, 16), wetland sediments (17), paddy soils (18–20), and land-freshwater interfaces (21). As an environment-friendly pathway, the anammox process has been applied in wastewater treatment systems (22, 23). These reports provide insight into the biological N cycle. The anammox process has been investigated in various natural and artificial habitats (24, 25).

The anammox process is performed by a deep-branching, monophyletic group of bacteria within the phylum Planctomycetes (26). To date, 16 species of anammox bacteria associated with five known candidate genera (Brocadia, Kuenenia, Anammoxoglobus, Jettenia, and Scalindua) have been described (27). The research on anammox is mostly focused on its role in the oceanic N cycle, and more than 50% of N₂ loss in some marine environments was found to be contributed by anammox (28, 29). However, recent studies have suggested that terrestrial soils possess an even higher diversity of anammox bacteria due to highly heterogeneous niches, which could provide sufficient oxic/anoxic interfaces (30, 31). Moreover, anammox activity was reported to account for 1 to 37% of the total N₂ loss from paddy soils (18, 19). The available evidence indicates that the anammox process plays an important role in the terrestrial N cycle.

Forests cover approximately 31.7% of the land globally and play an essential part in regulating global C and N cycling and global climates (32). Some forest soils may be favorable to the development of anammox bacteria and the occurrence of anammox due to both the nature of coexisting ammonium (NH₄⁺) and nitrate (NO₃⁻) and the presence of oxic/anoxic interfaces. Currently, there is only one study that detected the presence of the anammox bacterium “Candidatus Kuenenia” in a riparian forest wetland soil (33). However, the potential quantitative significance of anammox in forest soils has not yet been reported, and this information is indispensable to understanding the N cycling processes in forest ecosystems. In this study, we chose two temperate forest soils (i) to examine whether anammox was active and, if so, the importance of the anammox process as an N₂ producer and (ii) to detect the presence of anammox-related bacteria.

MATERIALS AND METHODS

Site description and soil sampling.

The study soils were collected from the Qingyuan Forest CERN (Chinese Ecosystem Research Network). The station is located in Qingyuan County of Liaoning Province, northeastern China (41°51′6.1″N, 124°54′32.6″E, 500 to 1,100 m above sea level) (34). The climate of this region is a continental temperate monsoon climate with an annual average precipitation of 811 mm (more than 80% falling during June to August) and an annual average temperature of 4.7°C (34). The growing season is from early April to late October. The brown forest soil belongs to Udalfs according to the second edition of U.S. Soil Taxonomy (35), and the soil texture is clay loam with 25.6% sand, 51.2% silt, and 23.2% clay (36).

Our sampling sites were located in a mixed broad-leaved forest (41°50′48″N, 124°56′01″E, 640 m above sea level) and a larch forest (41°50′58″N, 124°56′18″E, 625 m above sea level) with an age of 44 years. Overall, the sites had the same topographical features and were on soils developed from the same parental material, i.e., granite gneiss. In each forest, three 10-m by 20-m plots were laid out in July 2013. The plots of mixed broad-leaved forest consisted of Quercus mongolica, Juglans mandshurica, and Phellodendron amurense, including a small abundance of Larix olgensis and Fraxinus mandschurica. The plots of larch forest contain some shrubs, such as Acanthopanax senticosus, Acer tegmentosum, and Syringa wolfi.

Soil samples were collected from the mineral layer during the growing season in June of 2014. Five soil cores 2.5 cm in diameter were taken at random in each plot after removing the organic layer, and each soil core was separated into 0- to 10-cm, 10- to 20-cm, and 20- to 40-cm layers. The five soil cores of the same layer in each plot were mixed as one sample. Soil samples were placed in sterile plastic bags, sealed, and transported to the laboratory on ice. In the laboratory, soils were passed through a 2-mm sieve after roots and other visible debris were removed. The sieved soil samples were divided into three sets of subsamples. One set was stored at 4°C for ammonium (NH₄⁺) and nitrate (NO₃⁻) concentration analysis. The second part was used for isotope tracer incubation, and the remaining sample was air dried for C and N content and pH analyses.

¹⁵N tracer incubation experiments.

Slurry experiments were conducted to measure the potential rates of anammox and denitrification with the modified ¹⁵N isotope-tracing technique (3). Before the incubation experiments, freshly sieved soils were pooled and homogenized at the same layer of the three plots in each forest to form one composite sample. Six composite soil samples (0 to 10, 10 to 20, and 20 to 40 cm) from the two forests were used for the following slurry ¹⁵N tracer incubation experiments. Each composite sample was further subdivided into several laboratory replicates to obtain an average value for each sample.

Codenitrification can also produce ²⁹N₂ through the N-nitrosation of ¹⁵NO₂⁻ (produced from ¹⁵N-labeled NO₃⁻) with other non-¹⁵N-labeled N compounds, such as azide, salicylhydroxamic acid, and hydroxylamine (37, 38), when NO₃⁻ is ¹⁵N labeled. The first incubation experiment was therefore designed to quantify the combined potential rates of anammox and codenitrification. In this study, ultrahigh-purity N₂ gas (99.999%) was used to flush and fill the headspace to create anoxic conditions. Briefly, approximately 4 g of each composite fresh soil was weighed into 20-ml vials (Chromacol; 125 × 20-CV-P210) together with 3.5 ml of sterile deionized water. Then, the vials were shaken gently for several seconds and were capped tightly with gray butyl septa (Chromacol; 20-B3P; no.1132012634) and aluminum crimp seals (ANPEL Scientific Instrument Co. Ltd., Shanghai, China; 6G390150). The vials were further evacuated and flushed with ultrahigh-purity N₂ gas (100 ml min⁻¹) for 3 min. Following equilibration at atmospheric pressure, the vials were preincubated at 21°C in the dark for 24 h to remove residual oxygen (O₂) and to minimize NO_x⁻ (NO₂⁻ and NO₃⁻). After preincubation, each vial was again vacuumed and flushed with ultrahigh-purity N₂ gas using the method mentioned above. Subsequently, two treatments with five replicates were established: (i) sterile anoxic deionized water instead of tracer solution as a control and (ii) Na¹⁵NO₃ addition (¹⁵N at 99.26%; Shanghai Research Institute of Chemical Industry, China). The final concentration of ¹⁵NO₃⁻ was 100 μg of ¹⁵N g⁻¹ of fresh soil, which was obtained by injecting 0.5 ml of N₂-purged stock solution. The resulting soil slurries were stored for 24 h at 21°C in the dark. Biological activity was stopped by injecting 0.5 ml of 7 M ZnCl₂ solution into the incubation vials at the desired time. The headspace gas of each vial was sampled using a 1-ml gas-tight syringe for ¹⁵N₂ isotope analysis.

The second incubation experiment was conducted to confirm the presence of codenitrification and to distinguish the importance of anammox from codenitrification. Because relatively high rates of anammox plus codenitrification were observed in the mixed forest, only the soils from this forest were selected in this experiment. The soils were anoxically preincubated using the same method as described above, and the following treatments were performed with four replicates: (i) ¹⁵NH₄Cl (¹⁵N at 99.08%), (ii) ¹⁵NH₄NO₃ (¹⁵N at 99.14%), (iii) Na¹⁵NO₃ (¹⁵N at 99.26%), and (iv) Na¹⁵NO₂ (¹⁵N at 99.13%). The final concentrations of ¹⁵NH₄⁺, ¹⁵NO₃⁻, and ¹⁵NO₂⁻ were 100 μg of ¹⁵N g⁻¹ of fresh soil, achieved by injecting 0.5 ml of N₂-purged stock solution. The resulting soil slurries were incubated at 21°C in the dark. The headspace gas of each vial was sampled immediately using a 1-ml gas-tight syringe after adding tracer solution at 0, 5, 10, and 24 h for ¹⁵N₂ isotope analysis.

The third experiment was designed to determine the contribution of soil abiotic N₂ production. In this experiment, approximately 4 g of fresh soil from the 0- to 10-cm layer of the mixed forest was weighed into 20-ml vials, and soils were anoxically preincubated as in the first and second experiments. Then, the preincubated soils were autoclave sterilized twice in a 16-h period or were not sterilized for use as the control. Then, all vials were evacuated and flushed with ultrahigh-purity N₂ gas, as in the first experiment. Subsequently, 0.5 ml of sterilized ¹⁵N-labeled NaNO₂ or NaNO₃ solution was slowly injected into the vials. Each treatment was replicated four times. The amount of ¹⁵N added and the incubation conditions were the same as in the first and second experiments. Headspace gas was taken at 0, 5, 10, and 24 h after the onset of the experiment and was analyzed for ¹⁵N₂.

¹⁵N₂ analysis and rate calculation.

The ¹⁵N content of the N₂ in each 20-ml vial was determined by a continuous-flow isotope ratio mass spectrometer (IRMS) (IsoPrime 100 Isoprime Ltd., Cheadle Hulme, United Kingdom) at the Stable Isotope Ecology Laboratory of the Institute of Applied Ecology, CAS. The ¹⁵N₂ analysis system was built on the base of the IsoPrime trace gas analyzer (TG), as described by Yang et al. (39). We attached a separate section of tubing to connect the TG to a 12-port valve (model EC12WE; Valco Instruments Co. Inc., Houston, TX) for ¹⁵N₂ analysis and modified the scripts in IonVantage software and the setup of TG for manual injection analysis. The basic principle is to remove water vapor and trace gases, such as carbon monoxide (CO), interfering with ²⁹N₂ and ³⁰N₂ analysis via a cryo-trap and chemical traps and to separate N₂ and oxygen (O₂) in the gas samples utilizing a 5-Å molecular sieve column before introducing the gas into the IRMS. More details about the setup can be found in the work of Yang et al. (39). The levels of analytical precision for repeated manual analyses of ultrahigh-purity N₂ gas injection in our laboratory were 4.6 × 10⁻⁷ for R29 and 3.9 × 10⁻⁷ for R30; the coefficients of variation for these measurements were 0.0076% for R29 and 0.31% for R30 (n = 60). The precision determined from 28 manual injections of ultrahigh-purity N₂ gas over the course of a 48-h laboratory experiment was 0.037‰ for δ¹⁵N, comparable with those reported by Yang et al. (39). Therefore, in our study, we applied 0.11‰ δ¹⁵N or 4 × 10⁻⁵ atom% ¹⁵N, which is three times the precision for manual injections (39), to represent the minimum detectable change in ¹⁵N₂. Therefore, the minimum detectable N₂ flux is 0.009 nmol N g⁻¹ dry soil h⁻¹ over 24 h of incubation.

Headspace gas (600 μl) was sampled using a gas-tight syringe after the vials were shaken gently to equilibrate the gas between the dissolved and gaseous phases. Then, a 500-μl gas sample, after expulsion of 100 μl from the syringe to flush the dead volume (19 μl) of the needle, was manually injected into the N₂ sample loop (50 μl), and the areas for major (²⁸N₂), minor 1 (²⁹N₂), and minor 2 (³⁰N₂) from our IRMS as well as the ratios R29 (²⁹N₂/²⁸N₂) and R30 (³⁰N₂/²⁸N₂) were measured in both the enriched and reference samples (time zero).

According to the ratios R29 and R30 measured from the IRMS report, the mole fractions of ²⁹N₂ and ³⁰N₂ in the sample N₂ were calculated using equations 1 and 2 (39):

$$f^{29}=\frac{{{}^{29}\text{N}}_2}{{\text{N}}_2+{{}^{29}\text{N}}_2+{{}^{30}\text{N}}_2}=\frac{\frac{{{}^{29}\text{N}}_2}{{{}^{28}\text{N}}_2}}{1+\frac{{{}^{29}\text{N}}_2}{{{}^{28}\text{N}}_2}+\frac{{{}^{30}\text{N}}_2}{{{}^{28}\text{N}}_2}}=\frac{\text{R}29}{1+\text{R}29+\text{R}30}$$ (1)

$$f^{30}=\frac{{{}^{30}\text{N}}_2}{{{}^{28}\text{N}}_2+{{}^{29}\text{N}}_2+{{}^{30}\text{N}}_2}=\frac{\frac{{{}^{30}\text{N}}_2}{{{}^{28}\text{N}}_2}}{1+\frac{{{}^{29}\text{N}}_2}{{{}^{28}\text{N}}_2}+\frac{{{}^{30}\text{N}}_2}{{{}^{28}\text{N}}_2}}=\frac{\text{R30}}{1+\text{R}29+\text{R}30}$$ (2)

The production rates of ²⁹N₂ (P₂₉) and ³⁰N₂ (P₃₀) from each vial were estimated assuming that the vial headspace total N₂ concentration did not change during the 24-h incubation period (39). The mass of N₂ (M_total) in the vial headspace was calculated using equation 3:

$$M_{\text{total}}={\text{Density of N}}_2\times \text{volume}\hspace{0.17em}\text{of}\hspace{0.17em}\text{headspace}$$ (3)

The production rates of ²⁹N₂ (P₂₉) and ³⁰N₂ (P₃₀) in the vials were calculated using the following equations:

$$P_{29}=\{M_{\text{total}}\times [{(f^{29})}_{\text{t}}-{(f^{29})}_0]\}/(t\times M_{\text{soil}})$$ (4)

$$P_{30}=\{M_{\text{total}}\times [{(f^{30})}_{\text{t}}-{(f^{30})}_0]\}/(t\times M_{\text{soil}})$$ (5)

where t and 0 are the incubation time and time zero, respectively, and M_soil is the dry soil mass in the incubation vials (in grams).

In the ¹⁵NO₃⁻ incubation experiment, denitrification produces ²⁸N₂, ²⁹N₂, and ³⁰N₂, and both anammox and codenitrification produce ²⁸N₂ and ²⁹N₂ through random isotope pairing. In the single ¹⁵NO₃⁻ incubation, we can only separate the N₂ production rates by denitrification and by anammox plus codenitrification.

$$\begin{array}{l}{D}_{\text{total}}=D_{30}\times F_n^{-2},\hspace{0.17em}\hspace{0.17em}{D}_{29}=D_{\text{total}}\times 2\times (1-F_n)\times F_n,\\ D_{28}=D_{\text{total}}\times {(1-F_n)}^2\end{array}$$ (6)

where D_m is the production of N₂ through denitrification and F_n is the fraction of ¹⁵N in NO₃⁻.

³⁰N₂ production (P₃₀) is assumed to be solely from denitrification (i.e., P₃₀ = D₃₀). Rearranging equation 6 gives

$$\begin{array}{l}\begin{array}{ll}{D}_{\text{total}}=P_{30}\times F_n^{-2},\hfill & D_{29}=P_{30}\times 2\times (1-F_n)\times F_n^{-1}\hfill \end{array}\\ D_{28}=P_{30}\times F_n^{-2}\times {(1-F_n)}^2\end{array}$$ (7)

The ²⁸N₂ and ²⁹N₂ production rates by anammox plus codenitrification (AC) were then calculated, respectively, using the equations developed by Thamdrup and Dalsgaard for anammox (3):

$$\begin{array}{l}\begin{array}{ll}{\text{AC}}_{29}=P_{29}-D_{29},\hfill & {\text{AC}}_{28}={\text{AC}}_{29}\hfill \end{array}\times (1-F_n)\times F_n^{-1},\\ {\text{AC}}_{\text{total}}={\text{AC}}_{29}\times F_n^{-1}\end{array}$$ (8)

To further separate the contribution of ²⁹N production by anammox and by codenitrification, we have to rely on the results from the ¹⁵NH₄⁺ incubation. In the ¹⁵NH₄⁺ or ¹⁵NH₄⁺ and ¹⁴NO₃⁻ incubations under strict anaerobic conditions, we assumed that the ²⁹N₂ was produced solely from anammox (A₂₉). This assumption is reasonable because there would no ¹⁵NO₂⁻ production from the ¹⁵N-labeled NH₄⁺ due to a lack of oxygen, which is required in the first of step of NH₄⁺ oxidation in nitrification. The ²⁸N₂ from anammox (A₂₈) was calculated using the equation below:

$$\begin{array}{l}\begin{array}{lll}{A}_{28}=A_{29}\times F_a,\hfill & F_a=\frac{M_{{14}_{{\text{NH}}_4^{+}}}}{M_{{15}_{{\text{NH}}_4^{+}}}},\hfill & A_{\text{total}}=A_{29}+A_{28},\hfill \end{array}\\ A\%=\frac{A_{\text{total}}}{{\text{AC}}_{\text{total}}+D_{\text{total}}}\times 100\end{array}$$ (9)

where F_a is the ratio of the ¹⁴NH₄⁺ mass fraction to the ¹⁵NH₄⁺ mass fraction. Together with the results from the parallel ¹⁵NO₃⁻ incubation (AC₂₈ and AC₂₉), the ²⁹N₂ and ²⁸N₂ productions by codenitrification (C₂₉ and C₂₈) can be separated from those by anammox (A₂₉ and A₂₈) using the following equations:

$$\begin{array}{ll}{C}_{29}={\text{AC}}_{29}-A_{29},\hfill & C_{28}={\text{AC}}_{28}-A_{28},\hfill \\ C_{\text{total}}=C_{29}+C_{28},\hfill & C\%=\frac{C_{\text{total}}}{{\text{AC}}_{\text{total}}+D_{\text{total}}}\times 100\hfill \end{array}$$ (10)

Molecular detection of anammox bacteria.

Composite soil samples for isotope tracer and individual soil samples before compositing (22 soil samples) were used to detect the anammox bacteria. Total genomic DNA was extracted from 0.25 g of each soil sample using a PowerSoil DNA isolation kit according to the manufacturer's protocol.

A nested PCR assay was applied to amplify anammox 16S rRNA genes. The primer PLA46f-630r (40, 41) was used for the initial amplification of Planctomycetales 16S rRNA genes with a thermal profile of 96°C for 10 min, followed by 25 cycles of 60 s at 96°C, 1 min at 56°C, and 1 min at 72°C. The product from the first round of PCR was diluted 100-fold and subjected to the second round of amplification with the primer Amx368f-Amx820r (42) using a thermal profile of 96°C for 10 min, followed by 35 cycles of 30 s at 96°C, 1 min at 58°C, and 1 min at 72°C.

PCR products were verified for the correct amplicon size on a 1% agarose gel and were purified using a DNA purification kit (DP214, Tiangen Biotech, Beijing, China). The purified PCR products were ligated into pMD19-T vector (TaKaRa, Bio Inc., Shiga, Japan), and a clone library was constructed according to the manufacturer's instructions. Clones were randomly selected and verified for correct insertion of the DNA fragment by PCR with the universal primer set M13-47 and RV-M. Thirty positive clones from each clone library were randomly selected for sequencing. The retrieved sequences were checked using Chromas LITE (version 2.01) and were searched in the NCBI GenBank database using the BLAST program. The qualified sequences and their most closely related sequences obtained from the NCBI blast database were used for phylogenetic tree construction using MEGA 6.0 software by the neighbor-joining method (43) with the Kimura 2-parameter model. The sequences sharing more than 96% similarity were grouped into the same operational taxonomic unit (OTU). A bootstrap analysis with 1,000 replicates was applied to estimate the confidence values of the phylogenetic tree nodes.

Quantitative PCR.

The abundance of anammox bacteria was estimated by quantifying the hzsB gene with primer pair hzsB_396F (5′-ARGGHTGGGGHAGYTGGAAG-3′) and hzsB_742R (5′-GTYCCHACRTCATGVGTCTG-3′) (13). Quantitative PCRs were conducted in a 25-μl volume containing 12.5 μl of SYBR green premix (TaKaRa Bio Inc., Shiga, Japan), 0.6 μl of each primer (10 μM), and 1 μl of DNA template (1 to 20 ng μl⁻¹). The touchdown PCR was performed with a 5-min denaturation at 95°C, followed by 5 cycles of 1 min at 95°C, 64°C for 1 min (decrease of 1°C each cycle), 72°C for 45 s, and finally 72°C for 5 min. PCR product amplified with the same primer pair as mentioned above from the DNA template was cloned into a pMD 18-T vector (TaKaRa). Then, the positive clones were sequenced, and the plasmid DNA carrying the hzsB gene was extracted and purified. The concentration of the plasmid was then determined by NanoDrop, and the copy number of the hzsB gene per microliter was calculated. Tenfold serial dilutions of the plasmid were utilized as the standard for quantitative PCR.

Chemical analytical procedures for the soils.

Freshly sieved soils from each plot and composite soils for incubation were extracted with 2 M KCl solution. Extracted NH₄⁺ was measured through the indophenol blue method (>0.005 mg/liter [44]), and NO_x⁻ was determined by the copperized cadmium reduction method coupled with the modified Griess-I1osvay method (>0.01 mg/liter [44]). Soil total N and total carbon (C) contents were determined with an elemental analyzer (>10 μg of N and C; Elementar Analysen Systeme GmbH, Germany). Soil pH was determined in a 1:2.5 soil-water suspension. Soil bulk density samples were measured at each soil depth using a known-volume metal container. The gravimetric soil water content was calculated from the mass loss after drying for 24 h at 105°C.

Statistical analyses.

All reported values were expressed on a soil dry-weight basis. Statistical analyses were performed using SPSS software (version 16.0; SPSS Inc., Chicago, IL), including analysis of variance (ANOVA) and Pearson correlation analysis. Two-way ANOVA using soil layers and forest types as the main factor was used to differentiate the difference between forests. One-way ANOVA was also conducted to determine the difference between soil layers within each forest. Statistically significant differences were set at a P value of 0.05 unless otherwise stated.

Accession number(s).

The 16S rRNA gene sequences of anammox bacteria reported in this study have been deposited in the GenBank database under accession numbers KR088215 to KR088262.

RESULTS

Soil properties.

All of the examined soils of two forests were acidic, with pH values of 5.4 ± 0.1 to 6.1 ± 0.1 (Table 1). Total C and total N contents varied from 0.7% ± 0.1% to 5.0% ± 0.9% and 0.07% ± 0.02% to 0.44% ± 0.07%, respectively, and were approximately twice as high in the mixed forest soils as in the larch forest. The soil moisture ranged from 0.24 ± 0.01 to 0.47 ± 0.01 g of H₂O g⁻¹ of dry soil. The soil NH₄⁺ concentration varied from 2.5 ± 0.18 to 9.8 ± 0.18 mg of N kg⁻¹. The NO₃⁻ concentration ranged between 1.7 ± 0.05 and 7.0 ± 0.23 mg of N kg⁻¹, except for the 0- to 10-cm layer of the mixed forest, which had an especially high concentration, 25.5 ± 0.92 mg of N kg⁻¹. Among the soil layers, the surface soil had the highest NO₃⁻ concentration and total C and N contents (Table 1).

TABLE 1.

Soil properties for the two study forests^a

Soil layer (cm)	Soil bulk density (g cm−3)	Water content (g of H2O g−1 of dry soil)	pH	Total C (%)	Total N (%)	NH4+ (mg of N kg−1)	NO3− (mg of N kg−1)
Mixed forest
0–10	0.62 ± 0.04	0.47 ± 0.01	5.9 ± 0.07	5.0 ± 0.95	0.44 ± 0.07	2.5 ± 0.18	25.5 ± 0.92
10–20	0.87 ± 0.06	0.33 ± 0.01	5.8 ± 0.08	2.1 ± 0.46	0.23 ± 0.04	6.3 ± 0.05	7.0 ± 0.23
20–40	1.2 ± 0.10	0.26 ± 0.01	5.7 ± 0.10	1.0 ± 0.22	0.12 ± 0.02	5.8 ± 0.02	1.7 ± 0.05
Larch forest
0–10	0.91 ± 0.02	0.29 ± 0.01	5.4 ± 0.14	2.4 ± 0.37	0.23 ± 0.03	9.8 ± 0.18	5.3 ± 0.30
10–20	1.2 ± 0.09	0.24 ± 0.01	6.0 ± 0.09	1.0 ± 0.06	0.11 ± 0.01	6.2 ± 0.23	3.4 ± 0.19
20–40	1.1 ± 0.17	0.24 ± 0.01	6.1 ± 0.12	0.7 ± 0.15	0.07 ± 0.02	5.4 ± 0.45	2.0 ± 0.17

^aSoils were collected from three plots in each forest. The means ± standard errors (n = 3) are shown.

Potential anammox and codenitrification rate, contribution, and controlling factors.

Based on the ²⁹N₂ and ³⁰N₂ productions from the ¹⁵NO₃⁻ incubation in the first experiment, the combined potential rates of anammox and codenitrification and of denitrification were calculated for the examined forest soils. The potential activity of anammox and codenitrification in the mixed forest ranged from 0.07 ± 0.01 to 1.2 ± 0.18 nmol N g⁻¹ dry soil h⁻¹ (Fig. 1A), higher than that in the larch forest (0.01 ± 0.01 to 0.19 ± 0.02 nmol N g⁻¹ dry soil h⁻¹ [Fig. 1B]). The denitrification rates varied from 1.4 ± 0.05 to 7.1 ± 0.24 nmol N g⁻¹ dry soil h⁻¹ in the mixed forest and from 2.0 ± 0.11 to 3.4 ± 0.16 nmol N g⁻¹ dry soil h⁻¹ in the larch forest, which were both significantly greater than the corresponding combined rates of anammox and codenitrification (Fig. 1A and B). N₂ production in the 0- to 10-cm soil layer was significantly higher than in the deep soil layers (Fig. 1). The combined contribution of anammox and codenitrification to N₂ production was 4.9% ± 0.7% to 14.4% ± 1.8% in the mixed forest and 0.5% ± 0.5% to 5.5% ± 0.7% in the larch forest, with the remaining production attributed by denitrification (Fig. 1C). Soil water, total soil C and N concentration, and NO₃⁻ concentration were positively related with the N₂ production rates by anammox plus codenitrification and by denitrification (Fig. 2).

FIG 1.

Potential N₂ production rates and contributions over the 24 h of incubation after ¹⁵NO₃⁻ addition under anoxic conditions. Panels A and B show the results of the N₂ production rates in the mixed forest and the larch forest, respectively. Panel C indicates the combined contribution of anammox and codenitrification to total N₂ production. Different lowercase letters indicate a significant difference in the N₂ production rates among soil layers, and different uppercase letters indicate significant differences of N₂ production rates between denitrification and anammox plus codenitrification in each soil layer at the 0.05 level. Error bars represent standard errors (n = 5).

FIG 2.

Relationships between the rates of denitrification, anammox plus codenitrification, and soil variable characteristics. Linear regression is used to test the correlation. Adjusted R² values with the associated P values are shown when the correlation is statistically significant.

The ²⁹N₂ and ³⁰N₂ production rates and the contributions of three microbial processes for the soils collected from the mixed forest were further determined by ¹⁵NH₄⁺ labeling and ¹⁵NO₂⁻ labeling (Fig. 3). With ¹⁵NH₄⁺ addition only, soil ²⁹N₂ gas was detected only in the 0- to 10-cm layer, which increased almost linearly to 2.2 ± 0.11 nmol at 24 h (Fig. 3A). When amended with ¹⁴NO₃⁻, the levels of production of ²⁹N₂ gradually increased with time in all three soil layers and were 2.6 ± 0.04, 2.1 ± 0.07, and 2.1 ± 0.07 nmol at 24 h, respectively (Fig. 3B). These results suggest that NO₃⁻ was limited in the ²⁹N₂ production by the anammox reaction in the 10- to 20- and 20- to 40-cm soil layers due to consumption of NO₃⁻ in the preincubation before ¹⁵N labeling. ³⁰N₂ production was not observed in the ¹⁵NH₄⁺ addition alone or with ¹⁴NO₃⁻ during the incubation (Fig. 3A and B), confirming our assumption of no ¹⁵NO₂⁻ production from the ¹⁵N labeled NH₄⁺. Thus, the ²⁹N₂ from ¹⁵NH₄⁺ labeling incubation can be considered anammox production.

FIG 3.

Production of ²⁹N₂ and ³⁰N₂ over time in anoxic incubations of different soils from the mixed forest with the addition of ¹⁵NH₄⁺ (A), ¹⁵NH₄⁺ plus ¹⁴NO₃⁻ (B), ¹⁵NO₃⁻ (C), or ¹⁵NO₂⁻ (D). Each symbol represents the mean ± standard error (n = 4).

Compared with ¹⁵NH₄⁺ or ¹⁵NH₄⁺ and ¹⁴NO₃⁻ addition, the production of ³⁰N₂ and ²⁹N₂ with amended ¹⁵NO₃⁻ only significantly increased with time and decreased with soil depth (Fig. 3C). The amount of ³⁰N₂ at 24 h was highest in the 0- to 10-cm soil layer (221.0 ± 48.4 nmol) and was 5-, 6-, and 4-fold higher than ²⁹N₂ in the same 0- to 10-, 10- to 20-, and 20- to 40-cm soil layers, respectively (Fig. 3C), indicating the predominance of denitrification in N₂ production. When ¹⁵NO₂⁻ alone was added, ²⁹N₂ production was much larger than ³⁰N₂ production (Fig. 3D). The amount of ²⁹N₂ production (171.8 ± 4.1 to 214.5 ± 8.5 nmol) along the soil profile at 24 h was 12-, 29-, and 36-fold higher than for ³⁰N₂ (14.6 ± 0.9 to 6.0 ± 0.4 nmol), and ³⁰N₂ production was lower than with ¹⁵NO₃⁻ addition (Fig. 3C and D).

When adding ¹⁵NO₃⁻ alone to the sterilized soil (0 to 10 cm of soil of the mixed forest), ²⁹N₂ production at 24 h was detected at the small amount of 1.4 ± 0.8 nmol (only approximately 3% of the ²⁹N₂ production from unsterilized soil), and ³⁰N₂ production was not detected during incubation (Fig. 4B). These results demonstrated that the N₂ production from ¹⁵NO₃⁻ labeling incubation was predominately a biological process.

FIG 4.

Production of ²⁹N₂ and ³⁰N₂ in sterilized and unsterilized soil (0- to 10-cm mineral layer of the mixed forest) under anoxic incubation with the addition of ¹⁵NO₂⁻ (A) and ¹⁵NO₃⁻ (B). Each symbol represents the mean ± standard error (n = 4). Solid squares, ²⁹N₂ in unsterilized soil; dark gray squares, ²⁹N₂ in sterilized soil; open circles, ³⁰N₂ in unsterilized soil; light gray circles, ³⁰N₂ in sterilized soil.

A large level of ²⁹N₂ production was observed in the incubation of sterilized soil with ¹⁵NO₂⁻: up to 34.7 ± 2.4 nmol at 24 h, accounting for 22% of the production in unsterilized soil (158.1 ± 9.5 nmol) (Fig. 4A). ³⁰N₂ production was detected at 24 h in the unsterilized soil (0.71 ± 1.5 nmol) but was not found in the sterilized soil (Fig. 4A). These results from the incubation with ¹⁵NO₂⁻ further confirmed the existence of codenitrification in the N₂ production (i.e., microbial N-nitrosation of NO₂⁻ with unlabeled N compounds).

According to the results from the second and third experiments, the anammox rates in the mixed forest were 0.08 ± 0.001, 0.06 ± 0.001, and 0.06 ± 0.001 nmol N g⁻¹ dry soil h⁻¹ in the 0- to 10-, 10- to 20-, and 20- to 40-cm soil layers, respectively, accounting for 1.1% ± 0.2%, 2.5% ± 0.1%, and 6.6% ± 2.2% of the total N₂ production (Table 2). The codenitrification rates along the soil profile were 0.4 ± 0.1, 0.05 ± 0.02, and 0.1 ± 0.004 nmol N g⁻¹ dry soil h⁻¹, contributing 5.0% ± 0.7%, 1.8% ± 0.8%, and 12.4% ± 3.6% of total N₂ production, respectively (Table 2).

TABLE 2.

N₂ production rates and contribution to the total N₂ production of anammox, codenitrification, and denitrification along the soil profile in the mixed forest^a

Soil layer (cm)	Rate (nmol N g−1 dry soil h−1)			Contribution (%)
	Anammox	Codenitrification	Denitrification	Anammox	Codenitrification	Denitrification
0–10	0.08 ± 0.001 Ba	0.4 ± 0.1 Aa	7.5 ± 1.6 Aa	1.1 ± 0.2 Cb	5.0 ± 0.7 Bb	93.9 ± 0.8 Aa
10–20	0.06 ± 0.001 Bb	0.05 ± 0.02 Bb	2.3 ± 0.1 Ab	2.5 ± 0.1 Bb	1.8 ± 0.8 Bb	95.7 ± 0.7 Aa
20–40	0.06 ± 0.001 Bc	0.1 ± 0.004 Bb	0.9 ± 0.3 Ab	6.6 ± 2.2 Ba	12.4 ± 3.6 Ba	81.0 ± 5.9 Ab

^aDifferent lowercase letters indicate significant differences among soil layers, and different uppercase letters indicate significant differences among anammox, codenitrification, and denitrification in the same soil layer at the 0.05 level. Values are means ± standard errors (n = 4).

Detection of anammox bacteria.

Nested PCR of the 16S rRNA genes was conducted to detect anammox bacteria. The fragments with the targeting length (approximately 450 bp) were successfully amplified from all samples. In total, 763 sequences were retrieved from 22 libraries and were classified into 267 operational taxonomic units (OTUs) at a 97% cutoff. Thirty-seven OTUs (representing 89 sequences) were identified as belonging to the Planctomycetes phylum, whereas the remaining were assigned to Acidobacteria and were unclassified bacteria based on BLAST searches and phylogenetic analysis (Fig. 5; see also Fig. S1 in the supplemental material). Among the 37 Planctomycetes OTUs, two distinct OTUs were affiliated within the known anammox bacterial cluster, i.e., “Candidatus Brocadia fulgida” (from the 0- to 10-cm surface soil of the larch forest) and “Candidatus Jettenia asiatica” (from the 10- to 20-cm soil of the mixed forest) (Fig. 5). The remaining 35 Planctomycetes OTUs (41 sequences) diverged from the known anammox bacterial cluster in the phylogenetic tree.

FIG 5.

Phylogenetic tree of anammox and Planctomycetes bacterial 16S rRNA genes from forest soils amplified by nested-PCR primer sets Pla46F/1037R and Amx368F/Amx820R from larch forest and mixed forest soils. The obtained sequences in this study and reference sequences from GenBank were combined with the neighbor-joining method. Solid triangles indicate the OTUs detected in the present study; the numbers in parentheses indicate the number of sequences within each branch. The sequences obtained in this study are available in GenBank under accession numbers KR088215 to KR088262.

The amplification results showed that the PCR efficiency of standards was 91.0%, and the R² value of the standard curve was 0.994 (data not shown). The range of hzsB copy number of the standards was 2.3 × 10² to 2.3 × 10⁹. Although the reaction for the standards was successful, no amplification of the hzsB gene in our samples was detected (data not shown). PCR was also performed with the same procedure, but no amplification was observed after verification with a 1% agarose gel.

DISCUSSION

Presence of anammox.

Initially, we expected that the anammox process would occur in the upland forest soils where the oxic/anoxic interfaces were present and NH₄⁺ and NO₃⁻ coexisted. In the present study, we examined and estimated the potential anammox rates for two temperate upland forest soils through anoxic slurry incubations with ¹⁵N isotope tracers. Our results suggest that anammox occurs in the studied forest soils but at a very low rate and contribution to the total N₂ production.

With incubation of ¹⁵NH₄⁺ and ¹⁴NO₃⁻, we found detectable ²⁹N₂ production, although the amounts were small (2.1 to 2.6 nmol), for all three soil layers from the mixed forest (Fig. 3A and B). In the 10- to 20-cm and 20- to 40-cm mineral soil layers, ²⁹N₂ production was observed only when both ¹⁵NH₄⁺ and ¹⁴NO₃⁻ were added, indicating that ambient ¹⁴NO_x⁻ had been depleted in these two soil layers during preincubation. Furthermore, we found that ²⁹N₂ was produced at an increasing rate during incubation in the subsoils (Fig. 3B), which indicated that it took time for soil microbes to convert the ¹⁴NO₃⁻ added to ¹⁴NO₂⁻, which subsequently reacted with ¹⁵NH₄⁺ to form ²⁹N₂. Therefore, we can conclude that ²⁹N₂ production was a microbially mediated process.

In these ¹⁵NH₄⁺ or ¹⁵NH₄⁺ and ¹⁴NO₃⁻ incubations under strict anaerobic conditions, we can assume that ²⁹N₂ production is solely from anammox (A₂₉). This assumption is reasonable because there would be no ¹⁵NO₂⁻ production from the ¹⁵N-labeled NH₄⁺ due to a lack of oxygen, which is required in the first of step of NH₄⁺ oxidation in nitrification. This assumption is further supported by the lack of ³⁰N₂ production observed in this study (if ¹⁵NO₂⁻ was produced, ³⁰N₂ would be formed by denitrification). Anaerobic NH₄⁺ oxidation can occur with manganese oxide or iron oxide as an oxidant, producing either N₂ or NO₃⁻ (45, 46). Under denitrifying conditions, the NO₃⁻ would subsequently be reduced to N₂ or the NO₂⁻ produced via denitrification would be partially cometabolized with NH₄⁺ to form N₂ via codenitrification. However, these mechanisms cannot explain the oxidation of ¹⁵NH₄⁺ to ²⁹N₂ and the similar levels of ²⁹N₂ production between ¹⁵NH₄⁺ and ¹⁵NH₄⁺ plus ¹⁴NO₃⁻ incubation in this study, because manganese or iron leads to the formation of ³⁰N₂, and ²⁹N₂ production should increase with ¹⁵NH₄⁺ + ¹⁴NO₃⁻ incubation if codenitrification occurred, both of which were not observed (Fig. 3A and B).

According to equation 9 as given above, the anammox rate was 0.06 to 0.08 nmol N g⁻¹ dry soil h⁻¹ and accounted for 1% to 7% of the total N₂ production (Table 2), which was much lower than those in other environmental settings, e.g., water column, sediments, wetland soil, and paddy soils (see Table S1 in the supplemental material). However, the ²⁹N₂ production rates with ¹⁵NH₄⁺ incubations observed are higher than those observed for six agricultural soils in the United States (<0.02 nmol N g⁻¹ dry soil h⁻¹ [47]). Our results, together with those of Long et al., suggest that anammox is present but may be of minor importance in the total fixed N removal in upland soils.

The phylogenetic analysis of the 16S rRNA genes sequence further confirms the presence of the anammox process in the present study soils. In the present study, sequences closely related to “Candidatus Brocadia fulgida” and “Candidatus Jettenia asiatica” were detected, exclusively dominating the anammox bacteria community in the two studied forests, although nearly 90% of the sequences retrieved fell outside Planctomycetes clusters (Fig. 5; see also Fig. S1 in the supplemental material). However, the hzsB gene abundance was very low and was consistent with very low anammox rates (Fig. 3A and B; Table 2).

Co-occurrence of codenitrification in soil N₂ production.

Codenitrification is a microbial process producing N₂O and/or N₂ through the reduction of NO₂⁻ with other N compounds, including azide, NH₄⁺, salicylhydroxamic acid, and hydroxylamine (37, 38, 47). Codenitrification can occur in both fungi (e.g., Fusarium oxysporum) and bacteria (e.g., Streptomyces antibioticus) (37, 47, 48). This process with N₂ formation has previously been found in grassland and agricultural soils (47, 49, 50). The present study also demonstrated that codenitrification with N₂ formation coexisted with anammox and denitrification in N removal for forest soils under anaerobic conditions. With paralleling incubation of the same soil with both ¹⁵NH₄⁺ and ¹⁵NO₃⁻, as mentioned above, we are able to differentiate the contribution of codenitrification from anammox to ²⁹N₂. Our results for the soil from the mixed forest showed that codenitrification accounted for 2% to 12% of the total N₂ production (Table 2). These results were much lower than in previous studies for two pasture soils (92% and 97%, respectively [49, 50]) where the N₂ emission from codenitrification, including the part from anammox, was the dominant loss pathway. In addition, based on the original results from Long et al. (47) and our equations as given above, the contribution of codenitrification was recalculated to be 0 to 68% of the total N₂ production for the six agricultural soils in the United States (see Table S2 in the supplemental material), assuming that ²⁹N₂ production with ¹⁵NH₄⁺ addition was from anammox.

In the slurries amended with ¹⁵NO₃⁻, both ³⁰N₂ and ²⁹N₂ were significantly accumulated (Fig. 1 and 3C). The incubation with ¹⁵NO₃⁻ alone cannot differentiate the contribution of codenitrification from anammox to ²⁹N₂. The results from the first experiment thus showed the combined potential rates of codenitrification and anammox, which ranged between 0.01 and 1.2 nmol N g⁻¹ dry soil h⁻¹ and accounted for 12.4% and 2.8% of the total N₂ production in the mixed forest and in the larch forest, respectively, over the entire 40-cm soil depth (calculated using NO₃⁻ concentration weight [Fig. 1]). These results showed that anammox, at most, contributed 3% to 12% of the total N₂ production in the forest soils, which was comparable to the importance in many other environments (see Table S1). The importance of anammox would be overestimated if ignoring the presence of codenitrification in the case for our study (Table 2) and probably in some previous studies (see Table S1). Due to the atypical cell structure, anammox bacteria can use a broad range of organic and inorganic compounds as electron donors, e.g., hydroxylamine and methylamines (51, 52), to execute the anammox process, being similar to codenitrification. It is still difficult to clearly separate the anammox and codenitrification from each other through antibiotic methods. Our study has quantified the contribution of three processes (anammox, codenitrification, and denitrification) to N₂ production, using dual ¹⁵N labeling and ¹⁵N pairing techniques. Future studies will continue to focus on these processes in other forest soils.

Nitrite can chemically react with NH₂OH and azide to produce N₂O and/or N₂ (53). We investigated whether the ²⁹N₂ production observed in the first and second experiments was caused by abiotic reactions. We found that the ²⁹N₂ production in the sterilized soil amended with ¹⁵NO₂⁻ was 0.5 nmol N g⁻¹ dry soil h⁻¹, but its production accounted for 22% of that of the unsterilized soil (Fig. 4A). This result indicated that chemical reactions indeed existed, but the biological process was the main pathway in ²⁹N₂ production. With incubation of ¹⁵NO₂⁻, ²⁹N₂ production was much greater than ³⁰N₂ production in the unsterilized soil, opposite to the pattern of incubation with ¹⁵NO₃⁻ (Fig. 3C and D). These results highlighted that the NO₂⁻ added was more easily utilized by microbes with N compounds (e.g., NH₂OH, azide, and amino compounds) to form ²⁹N₂, e.g., through codenitrification, than utilized by denitrifiers to form ³⁰N₂ through denitrification. In addition, the ²⁹N₂ production with ¹⁵NO₂⁻ addition for the 0- to 10-cm mineral soil from the mixed forest, even after taking the abiotic production into account (1.8 nmol N g⁻¹ dry soil h⁻¹ [Fig. 4A]), was still much larger than the N₂ production with ¹⁵NH₄⁺ addition (0.06 nmol N g⁻¹ dry soil h⁻¹ [Fig. 3A]), indicating that processes other than anammox produced ²⁹N₂. With incubation of sterilized soil amended with ¹⁵NO₃⁻, neither ²⁹N₂ nor ³⁰N₂ was produced (Fig. 4B), demonstrating that N₂ was produced only after NO₃⁻ was biologically reduced to NO₂⁻.

Patterns along the soil profile.

Both the potential denitrification rates and the combined potential rate of anammox and codenitrification were higher in the 0- to 10-cm soil layers (Fig. 1 and 3; Table 2), which was similar to the case with previous studies with paddy soils (18, 19, 54). Their results indicated that higher NH₄⁺ concentration and NO_x⁻ concentration could stimulate higher anammox and denitrification activity. In the present study, we observed a higher NO₃⁻ concentration and organic matter and water contents in the 0- to 10-cm soil layer (Table 1), correlating with these potential rates (Fig. 2). It suggested that the NO₃⁻ concentration combined with organic matter and soil water may be the predominant factors for these microbial processes. Furthermore, we found that the contributions of anammox and codenitrification to total N₂ loss increased with increasing soil depth, and the contribution of denitrification decreased (Table 2), probably suggesting the lack of an available carbon source for denitrification in the deep layer. Although there were only 3 soil sampling points (0 to 10, 10 to 20, and 20 to 40 cm) in our study, this pattern was consistent with those studies of paddy soils (18, 20, 55) and natural wetland soil (56).

In summary, our results provided evidence for the presence of anammox in temperate upland forest soils. However, the anammox contribution to N₂ is minor, accounting for 1 to 7% of the total N₂ production. The anammox bacteria “Candidatus Brocadia fulgida” and “Candidatus Jettenia asiatica” were detected but in very low abundance, indicating the minor importance of anammox in N removal for the studied forest soils. In addition, we found that codenitrification was another potential N₂ emission pathway in forest soils, and its contribution to N removal (2 to 12%) was comparable to the contribution of anammox (1 to 7%). Furthermore, the importance of anammox and codenitrification increased with soil depth, whereas the importance of denitrification decreased with soil depth, suggesting that anammox and codenitrification communities may have different ecological niches than the denitrification communities. However, further research is required to examine the importance of anammox and codenitrification in other forest soils.