Supplement of Nitrification of archaeal ammonia oxidizers in a high-temperature hot spring

Abstract. The oxidation of ammonia by microbes has been shown to occur in diverse natural environments. However, the link of in situ nitrification activity to taxonomic identities of ammonia oxidizers in high-temperature environments remains poorly understood. Here, we studied in situ ammonia oxidation rates and the diversity of ammonia-oxidizing Archaea (AOA) in surface and bottom sediments at 77 °C in the Gongxiaoshe hot spring, Tengchong, Yunnan, China. The in situ ammonia oxidation rates measured by the 15N-NO3− pool dilution technique in the surface and bottom sediments were 4.80 and 5.30 nmol N g−1 h−1, respectively. Real-time quantitative polymerase chain reaction (qPCR) indicated that the archaeal 16S rRNA genes and amoA genes were present in the range of 0.128 to 1.96  ×  108 and 2.75 to 9.80  ×  105 gene copies g−1 sediment, respectively, while bacterial amoA was not detected. Phylogenetic analysis of 16S rRNA genes showed high sequence similarity to thermophilic Candidatus Nitrosocaldus yellowstonii, which represented the most abundant operational taxonomic units (OTU) in both surface and bottom sediments. The archaeal predominance was further supported by fluorescence in situ hybridization (FISH) visualization. The cell-specific rate of ammonia oxidation was estimated to range from 0.410 to 0.790 fmol N archaeal cell−1 h−1, higher than those in the two US Great Basin hot springs. These results suggest the importance of archaeal rather than bacterial ammonia oxidation in driving the nitrogen cycle in terrestrial geothermal environments.


Introduction
Nitrogen is a key element controlling the species composition, diversity, dynamics, and functioning of many ecosystems (Vitousek et al., 1997).Despite of recent processes in our understanding of nitrogen cycling activities in soils, fresh and marine waters, and sediments (Francis et al., 2005;He et al., 2007;Beman et al., 2008;Jia and Conrad, 2009;Konneke et al., 2005;Nicol and Schleper, 2006), gaps in knowledge associated with high-temperature ecosystems have prevailed (C.L. Zhang et al., 2008).Recently, some studies have elucidated nitrogen metabolism and cycling in high-temperature hot spring ecosystems (Dodsworth et al., 2011b;Nishizawa et al., 2013;Gerbl et al., 2014).In such systems, there has been evidence of microbial communities oxidizing ammonia, the first and rate-limiting step of nitrification (Reigstad et al., 2008;Hatzenpichler et al., 2008).Since the occurrence of a putative archaeal amoA gene in hot spring environments was first reported by Weidler et al. (2007) and Spear et al. (2007), Thaumarchaeota possessing ammonia monooxygenase (AMO) have been obtained from some terrestrial hot springs in the USA, China, and Russia (Pearson et al., 2008;C. L. Zhang et al., 2008).
Previous studies targeting ammonia oxidation in hot springs mainly focused on archaeal amoA gene (AOA) via a variety of culture-independent approaches (e.g., 16S rRNA clone library, biomarkers; Weidler et al., 2007;Francis et al., 2007;C. L. Zhang et al., 2008;Jiang et al., 2010;Xie et al., 2014).The results from these studies suggested that ammonia-oxidizing Archaea (AOA) may be ubiquitous in high-temperature environments and even more abundant than their bacterial counterparts, which has led to a hypothesis that Archaea rather than bacteria drive ammonia oxidation in high-temperature hot spring environments.This hypothesis, S. Chen et al.: Nitrification of archaeal ammonia oxidizers however, still needs to be verified.Currently, our knowledge about the activity of AOA in such high-temperature environments is largely constrained, especially due to the data deficiency of ammonia oxidation rates (Reigstad et al., 2008;Dodsworth et al., 2011b;Li et al., 2015).In situ incubation experiments are urgently required to verify the potential activity of AOA and their contribution to ammonia oxidation in such high-temperature environments.
In this study, we selected the Gongxiaoshe hot spring at Tengchong geothermal field as a representative site to test the hypothesis that Archaea rather than bacteria drive ammonia oxidation in high-temperature hot spring environments.The reasons for choosing the Gongxiaoshe hot spring as the research site are: (1) Ammonia concentration in the Gongxiaoshe hot spring water is 102.61 µg L −1 , thermodynamically favorable to ammonia oxidation (Shock et al., 2005); (2) Ammonia-oxidizing Archaea Candidatus Nitrosocaldus yellowstonii were dominant in the hot spring water and no AOB amoA genes were detected in the hot spring (Hou et al., 2013), indicating that the ammonia oxidation driven by Archaea might be active.Here, in combination of cultureindependent (fluorescence in situ hybridization, quantitative polymerase chain reaction (PCR), and clone library) and culture-dependent ( 15 N pool dilution technique) approaches, we provide direct evidence that AOA are indeed responsible for the major portion of ammonia oxidation in hightemperature hot spring environments.

Site description and chemical measurements
The Gongxiaoshe hot spring is a small pool with a diameter of ∼ 300 cm and a depth of ∼ 130 cm (Fig. 1).Hot spring water in the pool is well mixed and water chemistry shows no difference in different areas of the pool (G.Zhang et al., 2008).Sediments of the Gongxiaoshe hot spring are found to be only present at the margin of the pool and at the bottom of the pool, representing two typically sedimentary environments in this pool.The samples from the pool margins and sediments from the bottom of the spring, designated SS (surface sediments) and BS (bottom sediments), respectively, were collected using sterile equipment in April 2013.During transportation, all of the samples were packed with dry ice.They were then stored in a freezer at −80 • C in a lab for further analysis.
Temperature and pH were measured in situ in the hot water spring.Temperature was determined with an iButton thermometer (DS1922T, Dallas Semiconductor, USA).The pH was measured using a pH meter (SevenGo ™ pH meter SG2, Mettler Toledo, USA).Water samples for cation and anion analysis were filtered through a syringe filter with a 0.22 µm filtration membrane; these samples were diluted 10 times with deionized water and stored in 100 mL polypropylene bottles in the field because an analysis was carried out after 2 days.The cation concentrations were determined using an IRIS Advantage ICP-AES, whereas the anion (F − , SO 2− 4 , Cl − ) concentrations were determined using the Ion Chromatography System (DIONEX ICS-1500, Thermo Scientific, USA).The HCO − 3 concentration was measured using the Gran titration method (Appelo and Postma, 1996).The NH + 4 -N and NO − 3 -N concentrations were determined using a nutrient analyzer (Micromac-1000, Partech, UK).

15 N stable isotope tracing of nitrification activity
Gross N nitrification rates were determined in situ by the 15 N pool dilution technique.All of the nitrification measurements were conducted in 500 mL polycarbonate culture flasks (Nalgene) with a silicone plug that contained 400 mL of mud (∼ 1/3 sediment by volume).Two subsamples were collected from the bottom and surface sediments with 350 µL of K 15 NO 3 (485 µmol L −1 , at 10 % 15 N).For each sample, two experiments were conducted to measure the in situ nitrification activity: A1 (SS slurry + 15 NO − 3 ) and A2 (BS slurry + 15 NO − 3 ).Meanwhile, potential nitrification activity was determined in the presence of high ammonium concentration: B1 (SS slurry + 15 NO ).Two pairs of duplicate reactors were set up in four experiments.The reactors were incubated near the in situ conditions of the hot spring water at 77 • C for 30 and 120 min.At certain time intervals (e.g., 30, 120 min), 80 mL aliquots were collected from the experimental reactors with sterile serological pipettes and transferred to acid-cleaned 250 mL polypropylene bottles.Prior to filtration, 40 mL of KCl (3 M) was added to each sample bottle, and the samples were shaken at 120 rpm for 1 h and then centrifuged at 1600 × g for 10 min (Reigstad et al., 2008).The supernatant was filtered through a syringe filter containing a 0.22 µm filtration membrane; the supernatant was subsequently stored in acid-cleaned 60 mL polypropylene bottles at 4 • C, and analysis was performed after 2 days.
In the laboratory, the concentrations of NH + 4 and NO − 3 in the filtrate were determined by a nutrient analyzer (Micromac-1000, UK).The NO − 3 ( 15 NO − 3 and 14 NO − 3 ) ions of the filtrates were converted to N 2 O by denitrifying bacteria (Pseudomonas aureofaciens) lacking N 2 O reductase activity, and N 2 O was quantified by coupled gas chromatography isotope ratio mass spectrometry (GC-IRMS, Thermo Scientific, USA; Dodsworth et al., 2011a).The ammonia oxidation rates were calculated using the equations of Barraclough (1991) as were the concentrations and N isotope ratios of NO − 3 in the samples incubated for 30 and 120 min, respectively.

DNA extraction and purification
DNA was extracted by the SDS-based extraction method described by Zhou et al. (1996), with some modifications.Briefly, approximately 5 g samples were frozen with liquid nitrogen and milled three times.Then the powdered samples were mixed with 13.5 mL of DNA extraction buffer and 100 µL of proteinase K (10 mg mL −1 ) in tubes; these tubes were horizontally shaken at 225 rpm for 30 min at 37 • C.After shaking, 1.5 mL of 20 % SDS was added, and the samples were incubated in a water bath; the temperature of the water bath was maintained at 65 • C for 2 h.During this period, the tubes were subjected to gentle end-over-end inversions every 15 to 20 min.The supernatant fluids were collected after subjecting the tubes to centrifugation at 6000 × g for 10 min at room temperature; the collected supernatant tubes were subsequently transferred into 50 mL centrifuge tubes.The supernatant fluids were mixed with an equal volume of chloroform: isoamyl alcohol solution (24 : 1, v/v).The aqueous phase was recovered by centrifugation and precipitated with a 0.6 volume of isopropanol at room temperature; this process was carried out for at least 1 h.Crude nucleic acids were obtained by centrifugation at 16 000 × g for 20 min at room temperature; these crude nucleic acids were washed with cold 70 % ethanol and resuspended in sterile deionized water; the final volume of this solution was 100 µL.The crude nucleic acids were purified with a Cycle-Pure kit (Omega, USA).These crude nucleic acids were then resuspended in the elution buffer, and the final volume of the solution mixture was 50 µL; this solution was stored at −80 • C.

PCR and clone library construction
The 16S rRNA gene was amplified with purified genomic DNA as templates using universal primers.The primer pairs A21F (5 -TTC CGG TTG ATC CYG CCG GA-3 ) and A958R (5 -YCC GGC GTT GAM TCC AAT T-3 ) were chosen for Archaea (Delong, 1992) and Eubac27F (5 -AGA GTT TGA TCC TGG CTC AG-3 ) and Eubac1492R (5 -GGT TAC CTT GTT ACG ACT T-3 ) were chosen for bacteria (Lane, 1991).In a total volume of 50 µL, the reactions were performed using 1.25 U of Taq DNA polymerase (Takara, Japan).The amplification conditions were as follows: an initial denaturation was carried out at 94 • C for 4 min, and then, the same denaturation was continued at 94 • C for 1 min.Thereafter, annealing was carried out at 55 • C for 45 s, while extension was conducted at 72 • C for 60 s; the process was repeated for 30 cycles, followed by a final extension step at 72 • C for 10 min.The PCR products were excised after being separated by gel electrophoresis; a gel-extraction kit (Omega, USA) was used to purify the products in accordance with the manufacturer's instructions.The purified PCR products were cloned into pMD20-T vectors (Takara, Japan) and transformed into competent Escherichia coli DH5α cells.To select the positive clones, colony PCR was used to determine the presence of correctly sized inserts containing vector-specific primers M13f (5 -GTA AAA CGA CGG CCAG-3 ) and M13r (5 -CAG GAA ACA GCT ATGAC-3 ).

Sequencing and phylogenetic analysis
All of the clones were sequenced by the dideoxynucleotide chain-termination method.In this procedure, an ABI 3730 capillary electrophoresis sequencer (Applied Biosystem, Inc., USA) was coupled with the T vector universal primers M13f and M13r.The whole sequence of each clone was spliced using DNAMAN software (version 6.0), and the vector sequences were deleted; the presence of chimeras was checked using the Greengenes chimera check tool (Bellerophon server; Huber et al., 2004).The program DOTUR was used to determine the operation taxonomic units (OTU) for each sequence; 97 % similarity was considered as the cutoff for the chimeric sequences.To find closely related sequences in the GenBank and EMBL databases for phylogenetic analysis, none of the chimeric sequences were submitted to the Advanced BLAST search program.Phylogenetic trees were constructed using the neighbor-joining method and the software MEGA (version 5.05).A bootstrap analysis was used to provide confidence estimates of the tree topologies.Archaeal amoA gene fragments were amplified using the primer pair Arch-amoAF (5 -STA ATG GTC TGG CTT AGA CG-3 ) and Arch-amoAR (5 -GCG GCC ATC CAT CTG TAT GT-3 ; Francis et al., 2005).Bacterial amoA genes were also tested using the bacterial primer sets amoA 1F (5 -GGG GTT TCT ACT GGT GGT-3 ) and amoA 2R (5 -CCC CTC KGS AAA GCC TTC TTC-3 ; Rotthauwe et al., 1997).PCR cycling was performed by the method of Francis et al. (2005).In this method, PCR products from SS and BS were recovered from the gel slices using a gelextraction kit (Omega, USA) in accordance with the manufacturer's instructions.The purified PCR products from each type of sample were cloned into the pMD20-T vectors (Takara, Japan) and transformed into competent Escherichia coli DH5α cells.Cloning and sequencing were performed according to the above-mentioned process.There were 40-50 randomly selected colonies per sample analyzed for the presence of insert archaeal amoA gene sequences.

Quantification of 16S rRNA genes and amoA genes
Archaeal and bacterial populations were determined by quantifying their 16S rRNA genes with 344F-518R (Øvreas et al., 1998) and 518F-786R primer pairs (Muyzer et al., 1993), respectively.In addition, the abundance of AOA and AOB were quantified using amo196F-amo277R (Treusch et al., 2005) and amoA-1F and amoA-2R (Rotthauwe et al., 1997) primers, respectively.All sample and standard reactions were performed in triplicate.The SYBR Green I method was used for this analysis.The 20 µL reaction mixture contained 1 µL of template DNA (10 ng), a 0.15 µM concentration of each primer, and 10 µL of Power SYBR Green PCR master mix (Applied Biosystems Inc., USA); this reaction mixture was analyzed with ROX and SYBR Green I.The PCR conditions were as follows: 10 min at 50  (Wang et al., 2009).Melting curve analysis was performed after amplification, and the cycle threshold was set automatically using system 7500 software v2.0 Patch 6.The efficiencies of the qPCR runs were 87.8-105.6 % (R 2 = 0.992-0.999)for 16S rRNA genes and 102 % (R 2 = 0.998) for AOA.Primers targeting different genes are listed in Table 1.

Sample processing for FISH
To visualize Crenarchaea cells in situ, fluorescence in situ hybridization (FISH) was performed according to the procedure described by Orphan et al. (2002Orphan et al. ( , 2009)).Small aliquots of sediment were fixed overnight at 4 • C using 2 % formaldehyde in 1× PBS [145 mM NaCl, 1.4 mM NaH 2 PO 4 , 8 mM Na 2 HPO 4 (pH = 7.4)]; these aliquots of sediments were washed twice with 1× PBS and stored at −20 • C in ethanol: PBS (1 : 1, v/v) medium.The total supernatant was filtered through a polycarbonate filter (Millipore) under low vacuum (5 psi; 1 psi = 6.89 kPa).Filters were cut into suitably sized pieces and transferred onto untreated, round 2.54 cm glass slides.The transfer of filters onto glass slides was performed according to the procedure described by Murray et al. (1998).In this process, 5 µL of a 1× PBS solution was spotted onto a glass slide that was scored with a diamond pen prior to mapping, and half of the freshly prepared filter was used to invert the sample onto the slide; this inverted sample was then airdried.Prior to FISH, the samples on the glass slides were treated with an EtOH dehydration series (50, 75, and 100 % EtOH), dried, and stored at −20 • C. Hybridization and wash buffers were prepared according to the procedure described by Pernthaler et al. (2001).Here, 20 µL of hybridization buffer containing 35 or 20 % formamide was added to the samples on the glass slides.FITC-labeled oligonucleotide Cren679 probe described by Labrenz et al. (2010), was added to the hybridization buffer so that the final solution had a concentration of 5 ng µL −1 .
The hybridization mixtures on the slides were incubated for 1.5 h at 46 • C in a premoistened chamber.After hybridization, the slides were transferred into a preheated wash buffer and incubated for an additional 15 min at 48 • C. The samples were rinsed in distilled water and airdried in the dark.The microscopic images of the hybridized samples were recorded on a Leica Imager (Leica, DMI 4000B, Germany).

Nucleotide sequence accession numbers
The clone libraries for archaeal communities (21F-958R), bacterial communities (27F-1492R), and archaeal amoA genes (amoAF-amoAR) were constructed.All of the smallsubunit rRNA gene sequences and the amoA sequences were deposited in the GenBank/EMBL nucleotide sequence database under the following accession numbers: from KP784719 to KP784760 for partial 16S rRNA gene sequences and from KP994442 to KP994448 for the amoA sequences.

Ammonia oxidation rates
In the surface and bottom sediments (without NH + 4 stimulation), the near in situ rates of ammonia oxidation were estimated to be 4.80 ± 0.2 and 5.30 ± 0.5 nmol N g −1 h −1 using 15 N-NO − 3 pool dilution technique, respectively.In the meantime, the nitrate concentration increased from 2.84 ± 2 to 3.25 ± 2 µM in the surface sediments and from 2.33 ± 3 to 2.62 ± 3 µM in the bottom sediments, further providing evidence for strong nitrification activity under in situ conditions in the hot springs.Furthermore, the potential activity of ammonia oxidation was determined with ammonium amendments.The nitrate concentration increased significantly upon the addition of NH + 4 , and the ammonia oxidation rates recorded in the surface sediments and bottom sediments (with NH + 4 ) were 5.70 ± 0.6 and 7.10 ± 0.8 nmol N g −1 h −1 , respectively.

Archaeal community composition and phylogenetic analysis
A total of 152 archaeal clone sequences of 16S rRNA genes were obtained in this study.Phylogenetic analysis showed the distribution of the clone sequences into three monophyletic groups: Thaumarchaeota, Crenarchaeota, and Euryarchaeota (Fig. 4).In this study, the most abundant archaeal phylum was Thaumarchaeota.Among them, two phylotypes (SS-A19 and BS-A1) were the most dominant archaeal lineage, representing 89 and 86 % of the cloned archaeal sequences in surface and bottom sediments, respec-tively.These sequences were closely related to the thermophilic, autotrophic, ammonia-oxidizing archaeal Ca.N. yellowstonii (de la Tarre et al., 2008).The seven archaeal OTUs found here belonged to Crenarchaeota, which contains sequences recovered from hydrothermal vents and hot spring environments.In addition, two phylotypes (BS-A47 and BS-A8) that were branched with uncultured sequences belonged to Desulfurococcales, which was also recovered from sediments of the hot spring.Euryarchaeota also occurred in both the sediments, but with relatively low abundances.Phylotype BS-A80 is associated with Geoglobus ahangari, which belongs to Archaeoglobales and is capable of oxidizing organic acids (Kashefi et al., 2002).SS-A12, which represents four clones recovered from the surface sediments, showed 93 % similarity to an uncultured archaeal clone that was recovered from the Spring River.SS-A47 belonged to the Thermoplasmatales that were 96 % similar to their nearest neighbor sequence, which were collected from the Spring River.The other euryarchaeotal sequences BS-14 and BS-A80 were similar to their uncultured counterparts (from 96 to 99 % identity), which were mostly recovered from hightemperature geothermal environments.

Community analysis of AOA
A total of 113 archaeal amoA gene fragments were obtained from the two samples.They were all branched within the four distinct clusters of archaeal amoA sequences: cluster Nitrosopumilus, Nitrososphaera, Nitrosotalea, and Nitrosocaldus (Fig. 5).Cluster Nitrosotalea had one phylotype (SS-AOA-65) with 11 sequences (12 % of the total sequences).The closely related sequences in this cluster included characteristic crenarchaeotal group sequences that were obtained from alpine soil (with 98 % identity).Another clone, MX_3_OCT_18 (DQ501052), from estuary sediment was 96 % similar.

Quantitative PCR
The qPCR results (Fig. 2b) indicated that the abundance of the archaeal 16S rRNA gene in the two samples was similar, ranging from 1.28 to 1.96 × 10 7 gene copies g −1 of dry weight of sediments.However, the abundance of the bacterial 16S rRNA gene varied greatly, ranging from 6.86 × 10 6 to 4.25 × 10 8 gene copies g −1 of dry weight of sediments (Fig. S2 in the Supplement).The copy numbers of archaeal amoA genes in the surface and bottom sediments are 2.75 × 10 5 and 9.80 × 10 5 gene copies g −1 sediment, respectively.The copy numbers of the archaeal 16S rRNA genes in the bottom sediments were significantly higher than those of the bacterial 16S rRNA genes, with a ratio of 28.57.However, in surface sediments, the ratio of bacterial 16S rRNA genes to archaeal 16S rRNA genes is 3.32.

FISH
FISH was used to analyze the relative abundance of Crenarchaea in two samples.As expected, most metabolically active Crenarchaea cells and aggregated cells were detected by FISH probes (Cren679; Fig. 3).Based on the qPCR results, a high abundance of Crenarchaea in the hot spring sediments harbored amoA genes, providing strong evidence supporting the important role of Crenarchaea in the oxidation of ammonia.

Environmental factors affecting the occurrence of ammonia-oxidizing microorganisms
Temperature is likely a very important factor influencing microbial community structure.This interpretation is supported by the results of qPCR (Figs.2b and S2).The sediment samples from the bottom of pool (T = 77 • C) are dominated by Archaea, whereas the sediment samples from the margin of pool (T = 55 • C) are dominated by bacteria.In addition, no AOB were detected in both bottom and margin samples, indicating that it might be difficult for AOB to inhabit in hightemperature hot spring environments (Lebedeva et al., 2005;Hatzenpichler et al., 2008).Additionally, the abundance of AOA amoA gene in bottom sediments is slightly higher than that in margin sediments, reflecting that although AOA can adapt to a wide range of temperature, higher temperature could be more favorable to the growth of AOA (de la Torre, et al., 2008;Hatzenpichler et al., 2008;Jiang et al., 2010).Ammonia concentration may be another factor that influences the potential activity of AOA and AOB in hot springs.Because AMO in AOA has a much higher affinity for the substrate compared to a similar process in AOB, the ability of AOA to compete for ammonia in oligotrophic hot spring environments is also substantially higher than that of AOB (Hatzenpichler et al., 2008).In the Gongxiaoshe hot spring, the ammonia concentration of 102.61 µg L −1 is lower compared to other hot springs with high ammonia concentrations.This relatively low ammonia concentration may possibly be responsible for the absence of AOB in the Gongxiaoshe hot spring.

Composition and abundance of AOA
The rarefaction curves (Fig. S3) for archaeal 16S rRNA genes and amoA genes in the surface and bottom sediment samples reached a plateau, and their coverage values were relatively high (89-99 %).This result indicated that a large part of the archaeal/amoA diversity at this spring was probably included in the archaeal/amoA clone libraries.The ma- jority of archaeal sequences were closely related to Ca. N. yellowstonii, a known AOA, which may be responsible for the oxidation of ammonia in this spring.
In this study, phylogenetic analyses of archaeal amoA genes showed that Candidatus Nitrosocaldus yellowstonii dominated in both of the samples.This result also agreed with previous hot spring observations reported by Dodsworth et al. (2011b) and Hou et al. (2013).According to the sequences retrieved from NCBI, Nitrosotalea and Nitrososphaera clusters were closely related to the cluster soil.One possibility is that some of the amoA genes obtained in this study may derive from soil AOA, particularly those sequences in cluster Nitrosotalea and cluster Nitrososphaera, which have been widely found in sediments and soils.Those AOA from soil might have evolved multiple times and have adapted to high-temperature environments.Based on the analysis of the real-time PCR and FISH methods, our data indicate that the abundance of AOA is relatively high in both samples.The archaeal amoA gene copy numbers ranged from 2.75 to 9.80 × 10 5 per gram dry weight of sediments in this study.This is comparable to the abundance in other hot water springs [10 4 -10 5 copies g −1 (Dodsworth et al., 2011b)], but is lower than the abundance of the archaeal amoA gene in nonthermal environments, such as paddy rhizosphere soil [10 6 -10 7 copies g −1 (Chen et al., 2008)] and marine sediments [10 7 -10 8 copies g −1 (Park et al., 2008)].The bacterial amoA genes were not detected, indicating that AOB is absent or is a minority in this hot spring ecosystem.A predominance of archaeal amoA genes versus bacterial amoA genes indicated that ammonia oxidation may be due to the activity of Archaea in the Gongxiaoshe hot spring.

The role of AOA in the nitrification of terrestrial geothermal environments
In the surface and bottom sediments (without NH + 4 ), the ammonia oxidation rates calculated from the 15 N-NO − 3 pool di-lution data were 4.80 ± 0.2 and 5.30 ± 0.5 nmol N g −1 h −1 , respectively.The ammonia oxidation rates recorded in the surface sediments and bottom sediments (with NH + 4 ) were 5.70 ± 0.6 and 7.10 ± 0.8 nmol N g −1 h −1 , respectively.Moreover, the rates reported here were comparable with those observed in the two US Great Basin (GB) hot springs [5.50-8.60 nmol N g −1 h −1 (Dodsworth et al., 2011b)] and in two acidic (pH = 3, T = 85 • C) Iceland hot springs [2.80-7.00nmol NO − 3 g −1 h −1 (Reigstad et al., 2008)].However, the rates reported in this study were lower than those observed in some wetland sediments and agricultural soils [85-180 nmol N g −1 h −1 (White and Reddy, 2003;Booth et al., 2005)].
The ammonia oxidation rates in bottom sediments (without NH + 4 ) were slightly higher than those observed in surface sediments (without NH + 4 ).This result agrees with the distribution of archaeal amoA genes, which were found to be in higher abundance in the bottom sediment than in the surface sediment.High abundance of ammonia-oxidizing Archaea corresponds to high ammonia oxidation rates, which were consistent with the results reported by Isobe et al. (2012).Compared with the incubation experiments unamended with NH + 4 , the ammonia oxidation rate appeared to be stimulated after amendment with NH + 4 (1 M).There are indications that the ammonia concentration is an important factor affecting the rates of nitrification (Hatzenpichler et al., 2008).
To understand the relationship between the ammonia oxidation rates and abundances of amoA in the two samples, we specifically estimated the contribution of archaeal cells to nitrification.By assuming two amoA copies per cell (Bernander and Poplawski, 1997) and by comparing the ammonia oxidation rates with the qPCR results of AOA amoA per gram (however, some uncertainties of this method may still exist, with respect to the stage of cell cycle and the diversity of Archaea), the cell-specific nitrification rates were estimated to be 0.410 and 0.790 fmol N cell −1 h −1 in the surface and bottom sediments of the hot spring, respectively.These results are much higher than those for AOA in US hot springs [0.008-0.01fmol N cell −1 h −1 (Dodsworth et al., 2011b)].It is interesting that although the GB hot spring possesses higher amoA gene copies (3.50-3.90× 10 8 gene copies g −1 of dry weight) and higher NH + 4 concentration (663 µg L −1 ), it exhibits a lower cell-specific nitrification rate than the Gongxiaoshe hot spring.This may imply that both the abundance of AOA and the NH + 4 concentration are not important factors that control the cell-specific nitrification rates in high-temperature hot spring environments.The difference in cell-specific nitrification rates between the Gongxiaoshe hot spring and the GB hot spring may reflect the difference of AOA population structure in those two hot springs (Gubry-Rangin et al., 2011;Pester et al., 2012).In line with this AOA heterogeneity, cell-specific nitrification rates do not reflect the overall AOA abundance or NH + 4 concentration in these AOA-dominated hot springs.Alves et al. (2013) reported a similar case where soil dominated by AOA (clade A) exhibited the lowest nitrification rates, in spite of harboring the largest AOA populations.These results also suggest the importance of cultivation studies for comparative analysis of environmentally representative AOA in a wide variety of hot springs.

Conclusions
Combination of 15 N-NO − 3 pool dilution and molecular analyses demonstrate that the oxidation of ammonia by AOA occurs actively in the high-temperature the Gongxiaoshe geothermal system.The presence of considerable in situ nitrification rates in the hot spring is likely due to two dominant groups that include phylotypes that are closely related to the autotrophic AOA Ca.N. yellowstonii.The detection of archaeal amoA genes and the absence of AOB indicate that archaeal ammonia oxidizers, rather than AOB, significantly contribute to the nitrification in the Gongxiaoshe geothermal systems.Due to the AOA heterogeneity, cell-specific nitrification rates may not reflect the overall AOA abundance or NH + 4 concentration in the AOA-dominated hot springs.Our results shed light on the importance of AOA in driving the oxidation of ammonia in high-temperature environments, which may be ubiquitous in other terrestrial hot springs on Earth.
The Supplement related to this article is available online at doi:10.5194/bg-13-2051-2016-supplement.

Figure 1 .
Figure 1.The Gongxiaoshe hot spring, located in the Ruidian geothermal area.(a) A full view of the spring; (b) bottom sediments of the hot spring, designated BS; (c) an enlarged view of the white box from Fig. 1a, surface sediments of the hot spring; (d) surface sediments of the hot spring designated as SS; and (e, f) in situ nitrification activity and potential nitrification activity experiments in the field.

Figure 2 .
Figure 2. (a) Gross ammonia oxidation rates calculated from 15 N-NO − 3 pool dilution experiments on amended (added 14 NH + 4 ) or unamended SS and BS sediment slurries.This defines that the amendment with " 15 NO 3 " represents in situ nitrification activity, while 15 NO 3 plus 14 NH 4 is considered as potential nitrification activity.Bars represent the mean and standard error of the mean (n = 3) for 30 and 120 min incubation.(b) Abundance of archaeal 16S rRNA genes and archaeal amoA genes for SS and BS samples collected from the Gongxiaoshe hot spring.Data are expressed as gene copies per gram of sediment (dry weight).Error bars represent the standard deviation of the mean (n = 3).

Figure 3 .
Figure 3. Epifluorescence photomicrograph of Crenarchaeota cells and cell aggregates.(White and red arrows show the cells and carbonate crystals, respectively.Scale bar corresponds to 20 µm.)

Figure 4 .
Figure 4. Archaeal phylogenetic tree based on 16S rRNA gene sequences, including various 16S rRNA gene clones obtained from the Gongxiaoshe hot spring sediments (SS and BS) and cited some sequences from Hou et al. (2013; in red).The tree is constructed using the neighbor-joining method, and bootstrap confidence values over 50 % (1000 replicates) are shown.The scale bar represents the expected number of changes per nucleotide position.

Figure 5 .
Figure 5.The phylogenetic tree of archaeal amoA genes is cloned from the Gongxiaoshe hot spring sediments (SS and BS).The tree is constructed using the neighbor-joining method, and bootstrap confidence values over 50 % (1000 replicates) are shown.The scale bar represents the expected number of changes per nucleotide position.

Table 1 .
FISH probe and PCR primer pairs used in this study.