Accumulating evidence suggests that Thaumarchaeota may control nitrification in acidic soils. However, the composition of the thaumarchaeotal communities and their functioning is not well known. Therefore, we studied nitrification activity in relation to abundance and composition of Thaumarchaeota in an acidic red soil from China, using microcosms incubated with and without cellulose amendment. Cellulose was selected to simulate the input of crop residues used to increase soil fertility by local farming. Accumulation of -N was correlated with the growth of Thaumarchaeota as determined by qPCR of 16S rRNA and ammonia monooxygenase (amoA) genes. Both nitrification activity and thaumarchaeotal growth were inhibited by acetylene. They were also inhibited by cellulose amendment, possibly due to the depletion of ammonium by enhanced heterotrophic assimilation. These results indicated that growth of Thaumarchaeota was dependent on ammonia oxidation. The thaumarchaeotal 16S rRNA gene sequences in the red soil were dominated by a clade related to soil fosmid clone 29i4 within the group I.1b, which is widely distributed but so far uncultured. The archaeal amoA sequences were mainly related to the Nitrososphaera sister cluster. These observations suggest that fosmid clone 29i4 and Nitrososphaera sister cluster represent the same group of Thaumarchaeota and dominate ammonia oxidation in acidic red soil.
Thaumarchaeota are the most abundant archaea in the environment and may play an important role in ecosystem functioning. For example, they potentially contribute to climate change by producing greenhouse gases such as N2O (Santoro et al., 2011) or precursors of greenhouse gases such as CH4 (Metcalf et al., 2012) or sustain primary production by fixing CO2 (Tetu et al., 2013). All known ammonia-oxidizing archaea (AOA) are placed in the phylum Thaumarchaeota. Thaumarchaeotal ammonia oxidation consists of two steps, with NH3 firstly being oxidized to hydroxylamine by the enzyme ammonia monooxygenase (AMO). The mechanism of subsequent NH2OH oxidation to is still unclear (Vajrala et al., 2013). As AMO gene (amoA)-encoding Thaumarchaeota are ubiquitous and often more abundant than their bacterial counterpart, the relative contributions of and niche specialization between ammonia-oxidizing Thaumarchaeota and bacteria are presently attracting quite some research interest (Prosser & Nicol, 2008, 2012).
Archaeal ammonia oxidation appears to be dominant in acidic environments (He et al., 2012), as implicated by the increased relative abundance and activity of AOA vs. bacteria (AOB) (Nicol et al., 2008; Yao et al., 2011). Because ammonia (NH3) rather than ammonium () is the substrate of AMO, substrate availability is severely limited in acidic environments by protonation so that AOB are not active. By contrast, Thaumarchaeota possibly overcome the low concentrations of ammonia because of their unusually high substrate affinity (Martens-Habbena et al., 2009). Some Thaumarchaeota are capable of acidophilic growth (Lehtovirta-Morley et al., 2011), with mechanisms such as utilization of urea (Lu et al., 2012) or other mineralized organic N (Levičnik-Höfferle et al., 2012). Interestingly, in previous studies, growth, CO2 fixation, and ammonia oxidation activities of Thaumarchaeota in acidic soil microcosms were mostly related to the group I.1a (mainly Nitrosotalea) (Lehtovirta-Morley et al., 2011; Zhang et al., 2012; Lu & Jia, 2013), although other groups, such as I.1b, are also ubiquitous and sometimes even dominant in acidic soils (Gubry-Rangin et al., 2011; Hu et al., 2013). Whether these Thaumarchaeota are active as nitrifiers in acidic soils has not been well studied.
To increase the productivity of acidic soils, which are normally poor in nutrients, organic fertilizers such as crop residues and manure are often applied to the fields. Straw mainly consists of cellulose and hemicellulose, which can be degraded into various small organic molecules and may affect thaumarchaeotal ammonia oxidizers. There are some hints of organotrophic potential of Thaumarchaeota, although all known AOA can assimilate CO2 via a modified 3-hydroxypropionate/4-hydroxybutyrate carbon fixation pathway (Hatzenpichler, 2012). For example, AOA genomes encode genes of membrane transporters for organic compounds, implying uptake from the environment (Walker et al., 2010; Spang et al., 2012). Some organics, for example pyruvate and α-ketoglutarate, have been found to enhance the growth of Thaumarchaeota (Tourna et al., 2011), possibly explaining why they are often quite active in rhizosphere (Chen et al., 2008; Ke et al., 2013). Long-term application of organic fertilizer has also been found to result in increased abundance of Thaumarchaeota (Chan et al., 2013). However, the interaction between straw application and thaumarchaeotal ammonia oxidizers in acidic agricultural soils is still unclear.
Red soil is a major soil type in tropical and subtropical regions and is widely distributed in southern China. Previous surveys have revealed that the AOA community in acidic red soils could be affected by inorganic fertilization (He et al., 2007), soil management (Ying et al., 2010), or vegetation (Huang et al., 2012). Nonetheless, the relation between nitrification and thaumarchaeotal growth in red soils has rarely been studied (Zhang et al., 2012). In the present study, red soil microcosms with and without cellulose amendment were established, and molecular community analysis was performed to examine the dynamics of thaumarchaeotal ammonia oxidizers during the incubation. The objectives of this study were therefore to examine the thaumarchaeotal involvement in ammonia oxidization in red soil and to assess the effect of cellulose amendment.
Materials and methods
The soil was collected from an upland field (0–10 cm) near the Red Soil Ecology Experimental Station of CAS (28°15′N, 116°55′E) in Jiangxi Province, China. It can be classified as Ultisol in the USDA soil taxonomy and has a pH value of 5.0 (H2O). The collected soil was air-dried, well homogenized, and stored at 4 °C before use.
Soil microcosms were established with or without amendment of cellulose. Furthermore, inhibition treatments were run by adding acetylene to the headspace of the incubation bottles at a final partial pressure of 100 Pa (0.1%). As such, the experiment consisted of four treatment combinations of cellulose and C2H2 (−/−,−/+, +/−, +/+). Specifically, 7 g soil (DW) was added into 60-mL serum bottles, and 2 mg cellulose per gram soil consisting of 90% 12C- (Sigma-Aldrich) and 10% 13C-labeled cellulose (97 atom %, IsoLife, Wageningen, the Netherlands) was added to half of the microcosms. The 13C-cellulose was added to evaluate the consumption of cellulose from the 13CO2 release from the soil. After adjusting to 60% water-holding capacity (soil moisture about 18%), bottles were capped with butyl stoppers and statically incubated in the dark at 25 °C for up to 8 weeks. The headspace of the bottles was flushed weekly with CO2-free air to keep the microcosms aerobic. Concentration and isotopic ratio of CO2 in the headspaces were measured by gas chromatography (Shimadzu) and IRMS (Delta V Advantage, Thermo Scientific), respectively. Soil pH at the end of the incubation was measured with a microprocessor pH meter. All incubations were performed in triplicate, and destructive sampling for molecular community analysis took place at the 2nd, 4th, and 8th week.
Inorganic nitrogen measurement
Inorganic N including , , and was extracted from the soil with 2 M KCl at a soil/solution ratio of 1 : 2.5. After shaking at 200 r.p.m. for 1 h, the slurries were centrifuged at 10 000 g for 10 min. Ammonium in the supernatant was measured fluorometrically at an emission wavelength of 470 nm on a SAFIRE microplate reader (Tecan) as described previously (Murase et al., 2006). Nitrate and nitrite were analyzed using ion chromatography (Sykam).
DNA was extracted from 0.5 g soil with the FastDNA SPIN Kit for Soil (MP Biomedicals) following the manufacturer's instructions. The quantity of DNA was measured using a Qubit 2.0 fluorometer (Life technologies). Tenfold-diluted DNA did not significantly inhibit PCR; therefore, it was used in all downstream molecular analysis.
Quantitative PCR (qPCR)
qPCR was performed on an iCycler instrument (Bio-Rad). The primer sets of 771F/957R (Ochsenreiter et al., 2003) and CamoA-19f/CamoA-616r (Pester et al., 2012) were used for the quantification of thaumarchaeotal 16S rRNA and amoA genes, respectively, with the SYBR green-based reactions performed in triplicate for each sample. The qPCR standards were generated using plasmid DNA from representative clones containing the thaumarchaeotal 16S rRNA or amoA genes. A dilution series of the standard template across 7 orders of magnitude (101–107) for both genes was used per assay. The control was always run with water as the template instead of soil DNA extract. The qPCR amplification efficiencies for 16S rRNA gene were between 78.5% and 84.6%, with r2 values between 0.996 and 0.999; for thaumarchaeotal amoA qPCR assays, efficiencies were between 70.2% and 73.5%, and r2 values between 0.995 and 0.998 were obtained.
Terminal restriction fragment length polymorphism (T-RFLP)
T-RFLP was carried out as previously described with minor modification (Hernández et al., 2014) to determine the composition of thaumarchaeotal amoA genes. Briefly, PCR was performed using the primers of CamoA-19f/CamoA-616r mentioned above, with FAM (6-carboxyfluorescein) attached to the forward primer. The reactions were performed in 50 μL volume containing 0.2 mM dNTP each, 0.6 μM primers, 0.4 μg μL−1 BSA (bovine serum albumin), 2.5 units JumpStart Taq DNA polymerase (Sigma-Aldrich), and 1.0 μL 10-fold-diluted DNA. After purification, c. 100 ng PCR products was digested with five units NlaIII (New England Biolabs) in 20 μL volume at 37 °C for 3 h. Then, the reactions were desalted using SigmaSpin Post-Reaction Clean-up Columns (Sigma), and 2 μL of the desalted fragments was mixed with 11 μL Hi-Di™ formamide (Applied Biosystems) and 0.3 μL internal DNA standard MapMarker1000 (BioVentures), incubated for 3 min at 94 °C, and cooled on ice. Size separation was performed using a 3130 Genetic Analyzer (Applied Biosystems). The peaks with relative abundance of < 2% were excluded to avoid noisy background before calculating relative terminal restriction fragment (T-RF) abundances.
Cloning and sequencing
A 833-bp-long archaeal 16S rRNA gene fragment was amplified using the primer set A109F (Großkopf et al., 1998) and 957R (Ochsenreiter et al., 2003). The products of triplicate PCRs were pooled, purified, and cloned into the pGEM-T Easy Vector (Promega, Madison, WI). The recombined plasmid was transferred into E. coli JM109, and the positive clones were randomly selected and sequenced (GATC Biotech AG, Cologne, Germany). Putative chimeric sequences identified by Bellerophon (Huber et al., 2004) were discarded from the data sets. Operational taxonomic units (OTUs) were calculated using mothur (Schloss et al., 2009).
For the assignment of major T-RFs, clone libraries of thaumarchaeotal amoA genes were constructed following the same procedure described above using the CamoA-19f/CamoA-616r (Pester et al., 2012) primer set. The sequences were digested in silico and assigned to specific T-RF by comparison of fragment length.
After removing primer sequence, representative sequences for each OTU, their closest matches in the GenBank database searched by the blast program (http://blast.ncbi.nlm.nih.gov/Blast.cgi), as well as some reference sequences were used to calculate phylogenetic tree. Neighbor-joining trees and bootstrap support were calculated with mega version 4 software package (Tamura et al., 2007) based on the Jukes–Cantor model.
Pearson's correlation analyses were performed to assess the relationships among data sets. For multiple comparisons, one-way anova with Tukey's post hoc test was performed using the spss 13.0 package (SPSS), with α value of 0.05 selected for significance.
Nucleic acids accessories
Sequence data for thaumarchaeotal 16S rRNA and amoA genes were deposited with GenBank under accession numbers KF853585–KF853590.
Soil respiration and DNA yields
Microbial activity and growth during the incubation were assessed by soil respiration and DNA yields, respectively. In the control microcosms without cellulose amendment, CO2 production peaked in the first week and declined slowly during the incubation (Fig. 1a). Amendment with cellulose (2 mg g−1 soil) significantly stimulated respiration, resulting in approximately one order of magnitude higher CO2 release than without amendment. Due to the continuous decrease in respiration rate from the third week, as well as the slowdown of 13CO2 accumulation (data not shown), additional cellulose was added in the fifth week.
To avoid any effect of air-drying on DNA extraction, soil was rewetted for 12 h and was then used as the zero-time sample. For the control soil, the DNA yield increased from 0.24 to 0.76 μg g−1 soil during the initial 2 weeks incubation and remained stable afterward (Fig. 1b). Cellulose application significantly stimulated microbial growth, as suggested by the threefold higher DNA yield (2.56 μg g−1 soil) compared to the control microcosms after 8 weeks incubation.
Inorganic nitrogen dynamics
Nitrification activity in soil microcosms was calculated from the mineral nitrogen dynamics (Fig. 2). In the control microcosms, the net decrease in ammonium (8.5 μg -N g−1 soil) was similar to the increase in nitrate (8.3 μg -N g−1 soil), while nitrite was constantly low during the incubation. Assuming a linear production of during the incubation, the nitrification activity was about 0.14 μg -N d−1 g−1 soil. Both -N consumption and -N production could be completely inhibited by acetylene, an inhibitor of autotrophic nitrification. In addition, there was no apparent soil pH variation before and after incubation (data not shown). Cellulose enhanced the consumption of -N, and after incubation for 8 weeks, ammonium was below the detection limit (about 0.1 μg -N g−1 soil) despite the presence of acetylene. Meanwhile, there was no accumulation of nitrate corresponding to ammonium consumption. Only a marginal effect of C2H2 on mineral nitrogen was observed.
Abundance of thaumarchaeotal 16S rRNA and amoA genes
Population growth of Thaumarchaeota and amoA-encoding Thaumarchaeota during the incubation was assessed using qPCR with primers specific for 16S rRNA and amoA genes, respectively (Fig. 3). Significant increase was only observed for 16S rRNA gene in the microcosms without cellulose amendment, with the copies increasing from 2.78 × 105 to 2.11 × 106 g−1 soil. Meanwhile, the amoA genes increased from 6.54 × 104 to 1.12 × 105 copies g−1 soil. It should be noted that the copies of amoA genes were always less than those of 16S rRNA genes, as shown by the ratios of 16S rRNA:amoA genes ranging between 3.5 and 24.1. Significant positive correlations were observed between -N and 16S rRNA gene abundance (Fig. 4). In particular, the concomitant increase in -N and 16S rRNA gene copies could be well fitted by a linear regression (r2=0.752, P <0.001).
Thaumarchaeotal growth could be completely inhibited by addition of 100 Pa acetylene to the headspace of the incubation bottles (Fig. 3). There was also no significant increase in 16S rRNA and amoA gene abundances in all cellulose-amended microcosms. Bacterial amoA genes could not be detected in any of the microcosms using PCR amplification.
Community composition of Thaumarchaeota
Thaumarchaeotal composition was revealed by cloning and sequencing an 832-bp fragment of archaeal 16S rRNA genes. After removing putative chimeras, a total of 132 sequences with 20–24 sequences from each library were obtained, representing the thaumarchaeotal communities before and after 8 weeks of incubation. The sequences could be classified into three OTUs at 97% sequence identity, among which one (OTU_1) was especially abundant and accounted for 87–95% of the communities (Fig. 5). OTU_1 was highly similar to the soil fosmid clone 29i4 and formed a separate clade within the group I.1b of Thaumarchaeota (‘29i4 clade’ in Fig. 6) together with other sequences retrieved from the NCBI database. The other two rare OTUs accounting for 2–9% of the communities were most closely related to Nitrososphaera of group I.1b.
Restriction of amoA genes by Nla III produced two major T-RFs of 113 bp and 153 bp length, with the former being more dominant in the microcosms analyzed (Fig. 5). Sequence analysis showed that the 113-bp fragment shared 91% nucleic acid similarity with the closest matches in the NCBI database. This fragment as well as some environmental sequences fell within the Nitrososphaera sister cluster (Pester et al., 2012). Another fragment of 153 bp length belonged to the Nitrososphaera cluster.
The net nitrification rate of 0.14 μg -N d−1 g−1 soil in the absence of cellulose fell within the range measured in similar red soils (0.07–0.36 μg N d−1 g−1 soil) (Huang et al., 2012) and was tightly coupled with the growth of thaumarchaeotal populations. Concomitant inhibition of -N production and thaumarchaeotal growth by acetylene, together with the significant correlation between 16S rRNA gene abundance and -N concentration, provide compelling evidence for ammonia oxidation-dependent thaumarchaeotal growth in this acidic red soil. Assuming one ribosomal RNA operon for each thaumarchaeotal genome (Spang et al., 2012), the calculated cell-specific nitrification rate for Thaumarchaeota was 0.35 fmol -N cell−1 h−1, which is close to the calculated value for soil AOA N. viennensis (0.39 fmol -N cell−1 h−1) (Tourna et al., 2011). In addition, the availability of ammonia in this red soil was relatively low (< 0.4 μM at the determined pH, predicted by ; pKa = 9.25), while amoA-encoding Thaumarchaeota might be capable to exploit low concentrations of ammonia due to their unusual high affinity to ammonia (Martens-Habbena et al., 2009). It is therefore the Thaumarchaeota that were likely responsible for the ammonia oxidation observed under the experimental conditions. Moreover, we were unable to detect bacterial amoA genes using PCR amplification, indicating that abundance of AOB was consistently low in the soil during incubation.
Community analysis showed that the actively growing Thaumarchaeota in the red soil were mostly related to 16S rRNA gene fosmid clone 29i4 within group I.1b or to the Nitrososphaera sister cluster. The fosmid clone 29i4 is a 34-kb archaeal DNA fragment but does not contain an amoA homologue (Quaiser et al., 2002), while the Nitrososphaera sister cluster has been proposed based on its amoA genes (Pester et al., 2012). Links between the 29i4 clade and the Nitrososphaera sister cluster have previously been suggested (Nicol et al., 2008; Alves et al., 2013), and considering their dominance in the red soil 16S rRNA or amoA gene pool, both the 29i4 clade and the Nitrososphaera sister cluster likely represent the same group of Thaumarchaeota. Interestingly, the abundance of thaumarchaeotal 16S rRNA genes was consistently higher than that of the amoA genes. This could be hardly due to multiple copies of ribosomal RNA genes, because all known thaumarchaeotal genomes possess only one rRNA operon (Spang et al., 2012). Instead, it is possible that some group I.1b Thaumarchaeota lack amoA gene homologues, as having been found for the group I.1a thaumarchaeote Candidatus ‘Giganthauma sp.’ (Muller et al., 2010). Also, they could carry a modified amoA gene that cannot be amplified with the degenerate primers designed from the known thaumarchaeotal sequences (Pester et al., 2012), resulting in underestimation of the amoA gene abundance in qPCR assay. Lastly, qPCR-associated bias might have contributed to the differentiation in gene quantification.
Unlike previous studies, in which the nitrification activity in acidic soil was often linked to acidophilic Nitrosotalea (Zhang et al., 2010, 2012; Lu et al., 2012), our results suggest the activity of 29i4/Nitrososphaera sister cluster in the red soil. Despite its wide distribution and dominance in terrestrial environments (Hu et al., 2013), our knowledge about this group I.1b clade is surprisingly poor and even controversial. For example, there is evidence suggesting 29i4-related Thaumarchaeota are heterotrophs or mixotrophs as their growth was enhanced by addition of organic substrate (Xu et al., 2012). In another study, by contrast, enrichments of this cluster were found to be capable of growing under autotrophic conditions (Alves et al., 2013). One of our earlier studies also observed CO2 fixation by Nitrososphaera sister cluster Thaumarchaeota in a lake sediment using stable isotope probing (Wu et al., 2013). Future isolation of pure cultures and genomic analysis will facilitate to unravel their physiology and ecology.
Thaumarchaeotal growth was largely missing in cellulose-amended microcosms. This is unexpected because to our knowledge there is no previous evidence of cellulose being inhibitory to Thaumarchaeota or nitrification activity. The depletion of ammonium in these microcosms could not be attributed to nitrification, because of lacking -N accumulation and insensitivity to acetylene inhibition. Coupled nitrification–denitrification or anaerobic ammonium oxidation may contribute to the concomitant consumption of -N or -N; however, the most likely explanation for ammonium consumption is heterotrophic assimilation of mineral nitrogen by soil organisms (Verhagen et al., 1992), which was stimulated by cellulose amendment as implicated by the increased soil respiration rates as well as the increased DNA yields. The stimulation of ammonium assimilation by cellulose probably resulted in depletion of ammonium so that Thaumarchaeota could no longer proliferate in the acidic red soil.
In conclusion, our study indicated that in the acidic red soil, the growth of group I.1b Thaumarchaeota was dependent on ammonia oxidation and was sensitive to acetylene inhibition. A previously uncharacterized thaumarchaeotal clade that is closely related to soil fosmid clone 29i4 dominated the thaumarchaeotal community and probably controlled nitrification in the red soil.
We thank Melanie Klose and Peter Claus for technical assistance and Baozhan Wang and Yuji Jiang for the help in soil sampling. We thank also Marcela Hernández and Marc Dumont for helpful discussions. The study was financially supported by the Fonds der Chemischen Industrie, Deutschland.