ABSTRACT
The anaerobic oxidation of ammonium (anammox) process has been observed in diverse terrestrial ecosystems, while the contribution of anammox to N2 production in paddy soils is not well documented. In this study, the anammox activity and the abundance and diversity of anammox bacteria were investigated to assess the anammox potential of 12 typical paddy soils collected in southern China. Anammox bacteria related to “Candidatus Brocadia” and “Candidatus Kuenenia” and two novel unidentified clusters were detected, with “Candidatus Brocadia” comprising 50% of the anammox population. The prevalence of the anammox was confirmed by the quantitative PCR results based on hydrazine synthase (hzsB) genes, which showed that the abundance ranged from 1.16 × 104 to 9.65 × 104 copies per gram of dry weight. The anammox rates measured by the isotope-pairing technique ranged from 0.27 to 5.25 nmol N per gram of soil per hour in these paddy soils, which contributed 0.6 to 15% to soil N2 production. It is estimated that a total loss of 2.50 × 106 Mg N per year is linked to anammox in the paddy fields in southern China, which implied that ca. 10% of the applied ammonia fertilizers is lost via the anammox process. Anammox activity was significantly correlated with the abundance of hzsB genes, soil nitrate concentration, and C/N ratio. Additionally, ammonia concentration and pH were found to be significantly correlated with the anammox bacterial structure.
INTRODUCTION
The application of nitrogen-based fertilizers in agricultural fields causes various environmental problems, including eutrophication and habitat degradation. The fate of ammonium in the natural environments is of great importance for the microbial nitrogen cycle. For decades, heterotrophic denitrification was the only known pathway for nitrogen loss to the atmosphere. The discovery of the anaerobic oxidation of ammonium (anammox) coupled with nitrite reduction, with N2 as the end product in natural ecosystems, extends our understanding of the microbial diversity in the nitrogen cycle (1–3). Moreover, codenitrification also is involved in the removal of nitrogen from soils through the production of nitrous oxide (N2O) or dinitrogen gas (N2) (4). The accurate quantification of the processes removing fixed nitrogen is increasingly important to determine the fate of nitrogen in agricultural soils. The relative contribution of anammox and denitrification can be differentiated via 15N isotope-pairing techniques (1). However, codenitrification complicates any estimates of the contribution of anammox to N2 production, as it also can generate 29N2 by reducing 45N2O produced by using 14NH4+ and 15NO3−/15NO2− in 15N isotope-pairing experiments.
Anammox has been reported to account for >80% of total N2 production in the eastern tropical South Pacific oxygen minimum zone (OMZ) off northern Chile (5). Anammox bacteria were detected with broad biogeographic distribution in many naturally occurring anoxic environments, including aquatic ecosystems (6–9) and terrestrial ecosystems (10–15). However, their role related to activity in agricultural ecosystems remains poorly understood. Paddy fields represent one of the most significant N sinks in terrestrial ecosystems, yet a considerable proportion of N loss cannot be accounted for by the known pathways. Studies that performed investigations of the activity together with the diversity of anammox bacteria in paddy soils have been very limited (15, 16). Furthermore, the N loss due to anammox in paddy soil (Jiaxing, China) has been estimated only by Zhu and his colleagues (15). Hence, in the present study, in order to estimate the contribution of anammox to total N2 production on a larger scale, anammox rates of representative samples with diverse properties across southern China were measured, and the abundance and diversity of anammox-associated bacteria also were investigated.
The anammox process is mediated by bacteria affiliated to the order of “Candidatus Brocadiales,” of the phylum Planctomycetes (17), and is associated with five genera,“Candidatus Brocadia,” “Candidatus Kuenenia,” “Candidatus Anammoxoglobus,” “Candidatus Jettenia,” and “Candidatus Scalindua.” To date, all of these genera have been detected in paddy soils (14, 15), peat soils (18), and other terrestrial ecosystems (11, 19, 20). The anammox genotype is associated with hydrazine synthase, which consists of three subunits encoded by the genes hzsA, hzsB, and hzsC (21–23), responsible for the synthesis of hydrazine from nitric oxide and ammonium (22). The hzsB gene is a newly developed anammox phylomarker which has been successfully applied in the investigation of the biodiversity and quantitative analyses of anammox bacteria in various environments (9, 24, 25); hence, hzsB was used in this study.
China is the largest rice producer in the world, with 46 M ha in production (26), stably consuming ∼20 Tg of N-based chemical fertilizers per year (21.6 Tg N in 2005, 23.8 Tg N in 2011; China Statistical Yearbook [“Agriculture” section; http://www.stats.gov.cn/tjsj/ndsj/2012/indexeh.htm]). NH3 volatilization, N2O emission, runoff, and leaching result in substantial N loss from these systems (27), yet a significant proportion of N loss cannot be accounted for by these pathways alone (28). We hypothesize that anammox contributes significantly to this unaccounted loss; thus, the assessment of the anammox causing dinitrogen production would provide a robust basis for the development of potential management strategies for reducing N loss.
In the present study, 12 representative paddy soils covering 10 parent material types from 11 provinces in southern China were selected to: (i) evaluate the contribution of denitrification and anammox activity to N2 production using an 15N-tracer technique; (ii) determine anammox abundance via quantitative PCR (qPCR) of the hydrazine synthase (hzsB) gene and investigate the composition of anammox bacteria; and (iii) elucidate the impact of environmental variation on anammox activity under diverse environmental conditions.
MATERIALS AND METHODS
Soil sampling.A total of 12 paddy soils samples representing 10 parent material types from 11 provinces were collected from southern China (see Table S1 in the supplemental material). These sampling sites exhibited broad-scale variability in soil properties with respect to parent material, soil pH, and organic matter. Soil samples (0 to 20 cm) of four replicates (ca. 500 g each subsample) were collected during July 2013 from the dominant rice production fields in each province. The fields were grown with rice, but some sites (ZJ, AH, TY, and HB) had been drained off to different extents. For the fields which were still flooded, the overlaying water also was collected and kept on ice during transportation. Soil samples were placed in sterile plastic bags, sealed, and transported to the laboratory on ice. Each replicate sample was divided, with one subsample being incubated to determine denitrification and anammox activities immediately after arrival and another being processed through a 2.0-mm sieve for subsequent analysis of chemical components. The remainder was stored at −80°C for subsequent DNA extraction and molecular analysis.
Chemical analysis.Soil pH was determined in a 1:2.5 soil/water suspension. The total N and total C of the soil were measured by using dry combustion in a C/N instrument (Vario MAX C/N; Elementar, Germany). Soil organic C (SOC) was determined using a total organic carbon (TOC) analyzer (TOC-V CPH; Shimadzu, Japan). NO3−-N, NO2−-N, and NH4+-N were measured using an ion chromatograph (ICS-3000; Dionex) after extraction. The particle sizes of the soils were analyzed using a laser scattering particle analyzer (MS2000; Malvern, United Kingdom) after sieving through a 2-mm mesh to remove the gravel and plant roots. All analyses were performed in triplicate for each sample.
Measuring anammox and denitrification rates with 15N-labeled ammonium and nitrate.The presence and rate of anammox and denitrification was measured at in situ soil temperatures with a 15N-tracer technique (1, 29), with some modifications. Briefly, ∼3.5 g of soil (fresh weight) was transferred to 12.6-ml glass vials (Exetainer; Labco, High Wycombe, Buckinghamshire, United Kingdom), which then were filled with N2-purged in situ field water. For the fields which were not flooded, the incubation water was prepared in the laboratory by mixing 1 liter of double-distilled water (ddH2O) with 100 g fresh soil, followed by centrifugation (3,000 × g, 10 min). The supernatant was kept as the substitution for the in situ irrigation water. The resulting vials then were preincubated for 5 to 7 days to remove residual NOx− and oxygen according to our preliminary experiment when both NO2− and NO3− were under the detection limit of the ion chromatograph (0.05 to ∼0.1 ppm and 0.075 to ∼0.1 ppm for NO2− and NO3−, respectively). Subsequently, 100 μl of N2-purged stock solution of each isotopic mixture, that is, 15NH4+ [15N-(NH4)2SO4 at 99.14%], 15NH4+ + 14NO3−, and 15NO3− [15N-KNO3 at 98.15%], was injected through the stopper of each vial, resulting in a final concentration of about 100 μM N, which was added in excess based on the maximum total nitrogen (TN) concentration (ca. 60 μM N) detected in all collected samples. The incubations were performed at a temperature of 25°C ± 1°C. Time course incubations were carried out in duplicate (time points of 0, 3, 6, 12, and 24 h). ZnCl2 solution (200 μl, 7 M) was added at each time point during the incubation in order to stop microbial activity. The rate and potential contribution to N2 formation of either anammox or denitrification were calculated from the produced 29N2 and 30N2, measured by continuous-flow isotope ratio mass spectrometry (MAT253 with Gasbench II and autosampler [GC-PAL]; Bremen, Thermo Electron Corporation, Finnigan, Germany) as described before (1).
DNA extraction.DNA was extracted from approximately 0.25 g of each soil sample using a PowerSoil DNA isolation kit (MoBio Laboratories) in accordance with the manufacturer's instructions. DNA finally was eluted with 100 μl of the solution C6 (sterile elution butter), provided in this kit, and the DNA concentration and quality were measured using a NanoDrop ND-1000 (Thermo Scientific, Wilmington, DE). All DNA extracts were stored at −20°C for further diversity and abundance analysis.
Construction of clone libraries and phylogenetic analysis.Equal amounts of DNA extracted individually from four replicates pooled as a template for the nested PCR assay. A nested PCR approach was applied to amplify the 16S rRNA gene from anammox bacteria. The primer set PLA46F (a planctomycete-specific forward primer; 5′-GACTTGCATGCCTAATCC-3′) (30) and 630R (Escherichia coli positions 1529 to 1545; 5′-CAKAAAGGAGGTGATCC-3′) (31, 32) was used to amplify the Planctomycetales 16S rRNA genes with a thermal profile of 96°C for 10 min, followed by 35 cycles of 60 s at 96°C, 1 min at 56°C, and 1 min at 72°C. In the second round, the anammox bacterial 16S rRNA genes were amplified using the primer set Amx368F (5′-CCTTTCGGGCATTGCGAA-3′) (33) and Amx820R (5′-AAAACCCCTCTACTTAGTGCCC-3′) (34) with the following PCR conditions: 96°C for 10 min, followed by 25 cycles of 30 s at 96°C, 1 min at 58°C, and 1 min at 72°C (15).
PCR products were verified for the correct amplification on a 1% agarose gel and purified using the DNA purification kit (DP214; Tiangen Biotech, Beijing, China). The purified PCR products were cloned using the pMD19-T vector (TaKaRa, Bio Inc., Shiga, Japan) according to the manufacturer's instructions. Colonies were randomly picked up and verified for the correct insertion of the DNA fragment by PCR with universal primer set M13-47 (5′-CGCCAGGGTTTTCCCAGTCACGAC-3′) and RV-M (5′-GAGCGGATAACAATTTCACACAGG-3′). Forty positive clones from each sample were randomly selected for sequencing (MajorBio Ltd., Shanghai, China). The quality of the recovered sequences was checked using Chromas LITE (version 2.01) and subjected to a BLAST search in the NCBI database. All checked sequences were aligned using the Clustal W program. The phylogenetic analysis of the 16S rRNA gene sequences was conducted with MEGA 6.0 software using the neighbor-joining method. A bootstrap analysis with 1,000 replicates was applied to estimate the confidence values of the tree nodes. The distribution of the clones in different sites was visualized by using Circos software (35), and the Circos graphs were produced via Circos software online (http://circos.ca/).
Quantitative PCR.Thermal cycling and data analysis were carried out with a real-time PCR detection system (LightCycler 480; Roche) to assess the abundance of the anammox hzsB gene. The amplification was carried out in triplicate using the HSBeta396F-HSBeta742R primer (24) with a thermal profile of 3 min at 95°C, followed by 40 cycles of 30 s at 95°C, 30 s at 59°C, and 30 s at 72°C. The reaction mixture consisted of 20 ng template DNA, 0.3 μM each primer, 1× SYBR premix Ex Taq II, 1 mg ml−1 bovine serum albumin (BSA). Nontemplate controls were included in each amplification reaction. Standard plasmid carrying the anammox hzsB gene was generated by cloning the hzsB gene from samples as described above. Plasmid DNA containing the correct hzsB gene was extracted using a TIANprep mini plasmid kit (Tiangen Biotech, Beijing, China). The concentration of plasmid DNA was determined using a NanoDrop ND-1000 (Thermo Scientific, Wilmington, DE) for the calculation of hzsB gene copy numbers. A standard curve was obtained using 10-fold serial dilutions of standard plasmid DNA. A melting curve showed that only one peak at a melting temperature (Tm) of 84°C was detected, indicating the specificity of amplicons. Only the reactions with efficiencies between 90% and 110% were accepted.
Statistical analysis.Nonmetric multidimensional scaling (nMDS) was performed to cluster the different sites by their soil physicochemical similarities. Pearson correlation analyses were used to test the correlations among the anammox bacterial activity, abundance, and different environmental factors, using SPSS 18.0 (SPSS, Chicago, IL). The relationships between anammox activity and the environmental factors were determined by canonical correspondence analysis (CCA), using R software 2.14.0 with the vegan package. CCA-based variation partitioning analysis (VPA) was performed based on the diversity of the anammox OTUs using the correlation matrix.
RESULTS
Soil physicochemical properties.Seven of the soil samples (GZ, ZJ, CS, HB, GL, SC, and JX) maintained nearly neutral pH (6.40 to 7.36), while YT, TY, FZ, AH, and FST had significantly lower pHs (5.11 to 5.96; P < 0.05) (Table 1). The soil moisture (% dry weight) ranged from 44.43% to 66.96%, except for ZJ and SC, which were especially low at 35.08% and high at 79.17%, respectively. The concentration of NO3−-N and NH4+-N ranged from 0.3 to 5 mg kg−1 and 3 to 15 mg kg−1, respectively; thus, the equivalent nitrate and ammonium concentration ranged from 1.3 to 22.2 mg kg−1 and 3.9 to 19.5 mg kg−1, respectively. The soil organic matter concentrations varied from 13 to 46 g kg−1, while the total nitrogen concentrations ranged from 1.2 to 3.5 g kg−1, resulting in the C/N ratio ranging from 9.2 to 18.2. Among the 12 samples, GL showed the highest nitrate concentration, organic matter concentration, total nitrogen concentration, and C/N ratio. nMDS analysis also confirmed the high heterogeneity among the samples (see Fig. S1 in the supplemental material).
Physicochemical properties of the collected paddy soilsa
Anammox rate and contribution to N2 production in paddy soils.To determine anammox activity and its contribution to N2 production, homogenized soil incubations were performed at room temperature (∼25°C) using a 15N-nitrogen isotope-pairing technique. No significant accumulation of 15N2-labeled gas (29N2 and/or 30N2) was observed in any of the sample slurries amended with only 15NH4+ (see Fig. S2A in the supplemental material), indicating that all ambient 14NOx− had been depleted during preincubation. When both 15NH4+ and 14NO3− were added, 29N2 accumulated in each soil with no accumulation of 30N2 (see Fig. S2B). Therefore, as 30N2 can be produced only by denitrification, anammox must have been occurring. In slurries amended with 14NH4+ and 15NO3−, both 29N2 and 30N2 were significantly accumulated, as a result of both anammox and denitrification (see Fig. S2C). Anammox activity ranged from 0.27 ± 0.15 to 1.27 ± 0.16 nmol N g−1 dry paddy soil h−1, with the exception of GL, which exhibited the highest anammox rate (5.25 ± 1.83 nmol N g−1 dry paddy soil h−1) (Table 2). Denitrification rates varied substantially, from 4.21 ± 1.61 to 81.38 ± 22.32 nmol N g−1 dry paddy soil h−1 (significantly greater than the anammox rates); also, similar to its anammox rates, GL has the highest rate of denitrification activity (22.32 nmol N g−1 dry paddy soil h−1). The contribution to N2 production was calculated based on these rates, with anammox contributing between 0.76% (± 0.15) and 12.18% (± 2.67), with the remaining production attributed to denitrification (Fig. 1).
Rate of denitrification and anammox in different paddy soils
Contribution to N2 gas and abundance of anammox bacteria in paddy soils in southern China. The base map used is from the National Fundamental Geographic Information System of China. The red and green columns represent estimated anammox contribution to N2 production and hzsB gene abundance, respectively.
Abundance and composition of anammox bacteria.The abundance of anammox bacteria was estimated by qPCR of the hzsB gene (Fig. 1). The presence of the anammox genotype was confirmed by qPCR in all samples, with the abundance of hzsB ranging between 1.16 × 104 (± 0.29; site TY) and 9.65 × 104 (± 1.20; site GL) gene copies g−1 dry soil.
A total of 337 anammox bacterial 16S rRNA gene sequences were retrieved from the 12 paddy soil samples, which were clustered into 62 OTUs (97% nucleotide similarity). As expected, based on this targeted nested PCR analysis, all sequences from this study were affiliated with the family “Candidatus Brocadiaceae” (Fig. 2). The majority (52%) of OTUs were most closely related to “Candidatus Brocadia” (168/337), while only 4.5% were related to “Candidatus Kuenenia” (15/337). The remaining sequences formed two novel clusters, which may represent unidentified anammox genera and accounted for 24.3% (82/337) and 21.4% (72/337) of the anammox population, respectively. OTU5 and OTU47 were the most abundant (158/337), followed by OTU48, OTU46, and OTU54 (47/337). The diversity of anammox bacteria was highly variable across the 12 samples, with higher diversity in CS, YT, AH, and FZ (Shannon index of 1.84 to 2.71) and the lowest levels of diversity in GL and ZJ (Table 3). The distribution of each cluster in 12 paddy soils was visualized via Circos software (http://circos.ca/) (Fig. 3). “Candidatus Brocadia”-like sequences were detected in diverse samples, including JX, GZ, ZJ, HB, GL, and SC. All of the sequences in GL and ZJ belonged to “Candidatus Brocadia,” and more than 95% of sequences in HB also were affiliated with the genus of “Candidatus Brocadia.” The genus of “Candidatus Kuenenia” was only detected in the GZ sample. Cluster I was distributed in most of the collected samples, with the exception of GZ, ZJ, and GL, while cluster II was detected in CS, FST, AH, FZ, TY, and YT.
Neighbor-joining tree of phylogenetic analysis of the anammox bacterial 16S rRNA gene sequences using Kimura two-parameter distance in the MEGA 6.0 package. Clones with >97% sequence similarity were considered to be of the same OTU. Sequences of the reference known anammox bacteria are indicated in italics and followed by accession numbers. The sequences obtained in the present study are indicated in boldface and are identified by “OTU” followed by a number. The number in parentheses represents the number of clones in each OTU and the total number of clones. Bootstrap values (%) were generated from 1,000 replicates, and values of >50% are shown.
Alpha diversity analysis for anammox-targeted 16S rRNA genes from paddy soils
Distributions of each cluster in total anammox bacterial sequences in the 12 paddy soil samples. The data were visualized via Circos software (35) (http://circos.ca/). The length of the bars on the outer ring and the number on the inner ring represent the percentage of clones and the amount of clones in each sample, respectively. The bands with different colors show the source of each clone affiliated with different clusters.
The influence of physicochemical variables on anammox activity and diversity.The abundance of the hzsB gene, soil nitrate concentration, and C/N ratio were significantly correlated with the anammox activity in these samples (P < 0.01), while the soil nitrate concentration was significantly correlated with the abundance of the hzsB gene (P < 0.01) (Table 4). Anammox bacterial structure was found to be significantly correlated with ammonia concentration and pH (Fig. 4). Moreover, CCA-based VPA was performed to further separate the contribution of ammonia concentration, pH, and other properties on the anammox bacterial structure (Fig. 5). Ammonium concentration and pH explained a similar amount of variation of 8.76% and 8.25% in anammox bacterial diversity, respectively, and all other properties explained 55.71% of variance. Interactions between these factors contributed significantly more to the variation in anammox bacterial diversity (pH plus other properties, 65%; NH4+ plus other properties, 66.36%; pH plus NH4+, 21.55%), while only 21.28% remained unexplained.
Pearson correlation analyses of anammox rate and the soil properties, anammox diversity, and abundance in the collected paddy soils
Canonical correlation analysis (CCA) to show the relationship between anammox diversity and different environmental factors. Environmental variables were chosen based on significance calculated from individual CCA results and variance inflation factors (VIFs) calculated during CCA. The percentage of variation explained by each axis is shown, and two asterisks means the relation is significant (P < 0.01).
CCA-based VPA for anammox abundance. The percentage in the circle means the variation explained by the relevant factor, while the percentage on the line means the contribution explained by the interaction between these factors.
DISCUSSION
In the present study, we have shown that anammox activity can be detected across these diverse soil types, suggesting a ubiquitous distribution of anammox bacteria in paddy fields across southern China. The contribution of anammox to N loss was up to 15% in these selected paddy soils, demonstrating the important role of anammox in the global nitrogen cycling.
A recent study showed that anammox rates through a paddy soil depth profile (0 to 70 cm depth) with a high slurry manure fertilizer load ranged between 2.9 (0 to 10 cm depth) and 0.5 (60 to 70 cm depth) nmol N g soil−1 h−1 (15). A similar range (2.2 to 2.7 nmol N g soil−1 h−1) was observed in paddy soils from Japan, which received groundwater containing high nitrate concentrations derived from fertilizers and manure applied to the vegetable fields on the plateau above the paddy fields (16). Anammox rates observed in this study (0.27 ± 0.15 nmol N g−1 dry soil h−1 to 1.27 ± 0.16 nmol N g−1 dry soil h−1) exhibited a range similar to that of those studies, although a high rate was observed in GL (5.25 ± 1.83 nmol N g−1 dry paddy soil h−1), which may be due to higher concentrations of ammonium and nitrate than those of the other samples. In these water-logged ecosystems, oxygen availability and production of nitrite and/or nitrate may be the controlling factors for the anammox rate. Incubations with relatively high nitrate and ammonium (100 μM N) could stimulate anammox activity. Furthermore, we observed a wide range of anammox rates and anammox contribution to N2 production (0.6 to 15%), suggesting that anammox activity is highly variable in paddy soils. However, our paddy soils had significantly lower anammox contributions to N2 production than to other ecosystems, such as marine continental shelf sediment (20 to 80%) (1, 36), an anoxic water column in Golfo Dulce (19 to 35%) (2), which may be due to a large proportion of N2 produced by higher denitrification activity in our paddy soils.
Based on our results, N loss attributed to anammox was estimated to reach ∼2.5 Tg N per year on the basis of soil weight, rice cultivation area, and anammox rates obtained from slurry incubations. This suggests that ∼10% of applied N-based fertilizers (23.8 Tg N per year in 2011) is lost via the anammox process. Previous studies have assessed the fate of synthetic nitrogen fertilizer applied to the agricultural fields in China and demonstrated that only 35% of the N was utilized by the crops (28). N loss might be due to NH3 volatilization (11%), nitrification and denitrification (34%), leaching (2%), and N transport to ground and surface waters through runoff (5%), while the remainder (around 13%) was unknown (28). Our results suggest that the anammox process accounts for part of this unknown N loss. It is possible that slurry incubations overestimate the in situ anammox activity due to the following reasons. First, substrates are supplied in excess, which might enhance the anammox activity. Second, the prolonged preincubation might generate a favored condition for anammox rather than denitrification, since labile organic carbon could be consumed and exhausted, which also would overstate the contribution of anammox. Finally, codenitrification may be an additional pathway for N2 production in soils, which complicates any estimate of the contribution of anammox to N2 production, as codenitrification also can generate 29N2 by reducing 45N2O produced by using 14NH4+ and 15NO3−/15NO2− in 15N isotope-pairing experiments (37). Thus, in the present study, the anammox contribution might be overestimated through the 15N tracer technique in which the contribution of codenitrification is included. However, the ubiquitous presence of anammox bacteria and the abundance of the functional genotype provided robust evidence that anammox must play an important role in N2 production in these soils. Hence, more accurate quantifications of the anammox rates which can easily exclude the source of N2 between codenitrification and anammox are needed. The contributions of anammox, codenitrification, and denitrification to total N2 production in soil samples may be differentiated by using antibiotic inhibition coupled with 15N isotope-pairing techniques (37).
Anammox bacteria were detected in all of the soil samples based on the amplification of the anammox bacterial 16S rRNA genes, indicating the wide distribution of anammox bacteria in paddy soils, which would expand our knowledge of the distribution of anammox bacteria in environments, particularly in terrestrial ecosystems. In the present study, two known anammox genera were detected: “Candidatus Brocadia” and “Candidatus Kuenenia.” It is most likely that separate microniches for different anammox genera existed in the heterogeneous soils. Higher biodiversity observed in other agricultural ecosystems has been reported (10, 13, 15). However, a lower diversity was found in paddy soils from northeast China (14), where anammox bacterial communities consisted mainly of “Candidatus Scalindua” species. The “Candidatus Scalindua” species are best known to be dominant in marine and pristine environments, suggesting that the sampling sites have not been significantly affected by anthropogenic sources of pollution and urbanization (38–41). “Candidatus Scalindua” was not detected in any of the paddy soils in this study or in other fields in southern China (15), where the paddy soils were subjected to intense anthropogenic activity of long-term arable cultivation and human settlement. Compared to the homogeneous marine water column environment, the diverse distribution of anammox bacteria in terrestrial ecosystems shows higher variations and adaption to different soils. Our results showed that the most common anammox species in these soils were related to the genus “Candidatus Brocadia,” which has been observed in other paddy soils (13, 15). These results suggest that “Candidatus Brocadia” possess better adaptability in terrestrial ecosystems given that “Candidatus Brocadia” harbor diverse metabolism pathways (42); for example, they could use short-chain organic acids as alternative electron donors (43). Additionally, two novel clusters also were detected in our samples, suggesting that unknown anammox events are yet to be discovered and investigated.
Anammox activity, abundance, and diversity may be influenced by soil physicochemical properties, although how these variables influence the biogeography of anammox bacteria is poorly understood. We demonstrated that hzsB gene abundance, nitrate concentration, and C/N ratio were significantly correlated with the anammox activity. The abundance of anammox bacteria was estimated at a comparative level detected in various wetlands (12, 15). Most denitrifiers are capable of utilizing organic matter, resulting in releasing dissolved inorganic N (including NH4+ and NO2−) as the substrate for anammox bacteria. Anammox is favored in environments where NO3− is readily available (6, 44), and the availability of NO2− has been identified as the key determinant of anammox activity (45, 46). Therefore, NO3− can fuel anammox activity indirectly. However, NO2− is reliant on external sources, i.e., either NO3− reduction via denitrification (47, 48) or nitrification (7, 46). Of these, denitrification is thought to exert a crucial impact on anammox by regulating the amount of NO2−. Anammox bacteria can utilize NO2− only after its diffusion out of the cells of the denitrifiers, a situation that can occur when labile TOC (low C/N ratio) is limited and denitrification does not proceed to completion (7). Thus, the reactivity of available TOC regulates the extent to which denitrification reaches completion. The C/N ratio can be a proxy for the reactivity of soil organic matter, with low C/N content material generally representing fresh reactive organic matter (49, 50). This material favors denitrification over anammox, with the latter tending to dominate under high C/N conditions (7). Nonetheless, some studies illustrated that dissimilatory nitrate reduction to ammonia (DNRA) is the dominant pathway when levels of organic carbon input are high and nitrate levels are low (51, 52). Organisms using DNRA can outcompete denitrification at low NO3− due to their greater affinity for NO3−; thus, they can derive more energy from NO3− reduction than denitrifiers (53). However, in our incubation experiments, NO3− was not limiting, since the substrates were added in excess during anaerobic incubation. Ammonium concentration was identified to have a significant correlation with the alpha diversity of anammox bacteria. Ammonium is the main substrate for anammox bacteria, and nitrogen-rich fertilizers are the major source of soil ammonium in Chinese paddy soils. This implied that human activities (such as extensive N fertilization) indirectly influence the distribution of anammox bacteria in the environment (14, 54). As discussed previously, both “Candidatus Brocadia” and “Candidatus Kuenenia” commonly existed in the terrestrial ecosystem with human influence (11), which was in good agreement with our results.
In conclusion, our results indicate that anammox bacteria are ubiquitous in the paddy soils of southern China. The anammox community was dominated by the genus “Candidatus Brocadia,” and two unidentified clusters of anammox bacteria were detected. We demonstrated that the anammox process was responsible for the loss of a substantial quantity of N deriving from ammonia fertilizers. Anammox bacterial abundance, nitrate concentration, and C/N ratio were found to be the most important determinants for anammox rate, while ammonia concentration and pH were significantly correlated with the anammox bacterial composition. In a future study, quantitative characterization of in situ anammox rate and estimation of N loss via the anammox process at a finer spatial scale would provide a robust basis for the development of management practices that could be implemented to reduce soil N loss by anammox.
ACKNOWLEDGMENTS
This study was financially supported by the Natural Science Foundation of China (41090282), the Strategic Priority Research Program of Chinese Academy of Sciences (XDB15020300 and XDB15020400), and the International Science & Technology Cooperation Program of China (2011DFB91710). This work was supported in part by the U.S. Department of Energy under contract DE-AC02-06CH11357.
FOOTNOTES
- Received 14 August 2014.
- Accepted 16 November 2014.
- Accepted manuscript posted online 21 November 2014.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/AEM.02664-14.
- Copyright © 2015, American Society for Microbiology. All Rights Reserved.
